Help for package h2oplyr

Type:

Package

Title:

Tidy Data Manipulation on Remote 'H2O' Frames

Version:

0.2.6

Description:

Provides selection, filtering, mutation, aggregation, joins, reshaping and window operations on remote 'H2O' frames. Data manipulation is delegated to 'H2O'; no local table backend is used. Includes an English tutorial and reproducible examples using Palmer penguin measurements.

License:

MIT + file LICENSE

Encoding:

UTF-8

Depends:

R (≥ 4.1.0)

Imports:

h2o (≥ 3.44.0.3)

Suggests:

knitr, rmarkdown

VignetteBuilder:

knitr

SystemRequirements:

Java (>= 8), a running H2O cluster matching the h2o R package version

NeedsCompilation:

Packaged:

2026-09-29 07:07:52 UTC; carlosortega

Author:

Carlos Ortega [aut, cre]

Maintainer:

Carlos Ortega <cof@qualityexcellence.es>

Repository:

CRAN

Date/Publication:

2026-10-15 15:20:15 UTC

Remote H2O Data Manipulation

Description

Tidy-style manipulation of H2O frames using H2O as the only data backend. Start a matching H2O cluster explicitly, then use h_penguins() for the included tutorial data.

H2O %.%

Description

Usage

`%.%`(lhs, rhs)

Arguments

lhs

The left-hand pipeline value.

rhs

A function call receiving lhs as its first argument.

Value

The result of the right-hand function call.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      rows(x %.% h_filter(y > 30), 3)
  }
}

H2O all of

Description

Usage

all_of(x)

Arguments

x

An H2OFrame, an h2o_tbl, or (for selection helpers) a character pattern or vector. Local tables are not accepted.

Value

A column-selection specification containing metadata only.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      cols(h_select(x, all_of(c("id", "grp"))), c("id", "grp"))
  }
}

H2O any of

Description

Usage

any_of(x)

Arguments

x

An H2OFrame, an h2o_tbl, or (for selection helpers) a character pattern or vector. Local tables are not accepted.

Value

A column-selection specification containing metadata only.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      cols(h_select(x, any_of(c("id", "absent"))), "id")
  }
}

Access the remote H2O frame

Description

Usage

as_h2o(x)

Arguments

x

An H2OFrame, an h2o_tbl, or (for selection helpers) a character pattern or vector. Local tables are not accepted.

Value

An H2OFrame expression (vector helpers return one column).

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      stopifnot(inherits(as_h2o(x), "H2OFrame"))
  }
}

H2O contains

Description

Usage

contains(x)

Arguments

x

An H2OFrame, an h2o_tbl, or (for selection helpers) a character pattern or vector. Local tables are not accepted.

Value

A column-selection specification containing metadata only.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      cols(h_select(p, contains("bill")), c("bill_length_mm", "bill_depth_mm"))
  }
}

H2O desc

Description

Usage

desc(x)

Arguments

x

An H2OFrame, an h2o_tbl, or (for selection helpers) a character pattern or vector. Local tables are not accepted.

Value

Its input, unchanged; interpreted specially by h_arrange.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      z <- h_arrange(x, desc(y))
      equal_sum(as_h2o(h_slice_head(z, 1))[, "y"], 60)
  }
}

H2O ends with

Description

Usage

ends_with(x)

Arguments

x

An H2OFrame, an h2o_tbl, or (for selection helpers) a character pattern or vector. Local tables are not accepted.

Value

A column-selection specification containing metadata only.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      cols(h_select(x, ends_with("1")), "x1")
  }
}

H2O everything

Description

Usage

everything()

Value

A column-selection specification containing metadata only.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      cols(h_select(x, everything()), names(as_h2o(x)))
  }
}

Check an existing H2O connection

Description

Usage

h2o_available()

Value

A logical scalar.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      stopifnot(h2o_available())
  }
}

Wrap an existing remote H2O frame

Description

Usage

h2o_tbl(x)

Arguments

x

An H2OFrame, an h2o_tbl, or (for selection helpers) a character pattern or vector. Local tables are not accepted.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      rows(h2o_tbl(as_h2o(x)), 6)
      stopifnot(all(dim(x) == c(6, 7)), nrow(x) == 6, ncol(x) == 7, identical(names(x), 
          names(as_h2o(x))))
  }
}

H2O across

Description

Row data remain in H2O. Base R handles expressions, column names, configuration and remote object handles. H2O may return schema metadata or scalar reductions to the client; no local table transformation is performed. Mutation is ungrouped. Named expressions are evaluated sequentially. The closed expression interpreter accepts arithmetic, comparisons, is.na, membership in literal vectors, supported H2O math/text conversions and h2oplyr vector helpers. Unknown functions and other package namespaces are rejected. One-argument across closures are interpreted, never executed over downloaded rows.

Usage

h_across(.cols, .fns, .names = "{col}_{fn}")

Arguments

.cols

A column selection or character vector of column names.

.fns

A function or named list. Summary functions must be supported aggregates; mutation functions must have one argument and a supported expression body.

.names

Output naming template using col and fn tokens enclosed in braces.

Value

An across specification interpreted by mutation or summary verbs.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      z <- h_mutate(x, h_across(c("x1", "x2"), list(double = function(z) z * 2)))
      equal_sum(as_h2o(z)[, "x1_double"], 114)
  }
}

H2O anti join

Description

Row data remain in H2O. Base R handles expressions, column names, configuration and remote object handles. H2O may return schema metadata or scalar reductions to the client; no local table transformation is performed. Equijoins use H2O merge. Missing-key behavior follows H2O, not SQL NULL semantics. Output row order is unspecified; sort explicitly. Full joins combine a left join and unmatched right rows. Character keys are converted to categorical keys remotely. Suffix collisions and violated relationship constraints raise errors.

Usage

h_anti_join(x, y, by = NULL)

Arguments

x

An H2OFrame, an h2o_tbl, or (for selection helpers) a character pattern or vector. Local tables are not accepted.

y

A second remote table for joins, or a scalar comparison value for h_na_if.

by

Join column names; a named character vector maps left names to right names. For h_count, an optional grouping selection.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      rows(h_anti_join(h_slice_head(x, 3), h_slice_tail(x, 4), by = "id"), 2)
  }
}

H2O arrange

Description

Usage

h_arrange(.data, ..., na_last = TRUE)

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

...

Column selections, named expressions, remote tables or conditions, as appropriate to the verb. See the example and details.

na_last

Whether missing sort keys follow nonmissing values.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      z <- h_arrange(x, grp, desc(y))
      equal_sum(as_h2o(h_slice_head(z, 1))[, "y"], 20)
  }
}

H2O between

Description

Usage

h_between(x, left, right)

Arguments

x

An H2OFrame, an h2o_tbl, or (for selection helpers) a character pattern or vector. Local tables are not accepted.

left

Inclusive lower bound.

right

Inclusive upper bound.

Value

An H2OFrame expression (vector helpers return one column).

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      equal_sum(h_between(v, 20, 40), 3)
  }
}

H2O bind cols

Description

Usage

h_bind_cols(...)

Arguments

...

Column selections, named expressions, remote tables or conditions, as appropriate to the verb. See the example and details.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      cols(h_bind_cols(h_select(x, id), h_select(x, y)), c("id", "y"))
  }
}

H2O bind rows

Description

Row data remain in H2O. Base R handles expressions, column names, configuration and remote object handles. H2O may return schema metadata or scalar reductions to the client; no local table transformation is performed. Inputs must have the same column-name set and compatible H2O types; column order is aligned. Missing-column union is not implemented.

Usage

h_bind_rows(...)

Arguments

...

Column selections, named expressions, remote tables or conditions, as appropriate to the verb. See the example and details.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      rows(h_bind_rows(x, x), 12)
  }
}

H2O case when

Description

Usage

h_case_when(..., .default = NA)

Arguments

...

Column selections, named expressions, remote tables or conditions, as appropriate to the verb. See the example and details.

.default

Default scalar or compatible H2O column when no condition matches.

Value

An H2OFrame expression (vector helpers return one column).

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      z <- h_mutate(x, label = h_case_when(y >= 40 ~ 2, TRUE ~ 1))
      equal_sum(as_h2o(z)[, "label"], 9)
  }
}

H2O coalesce

Description

Usage

h_coalesce(...)

Arguments

...

Column selections, named expressions, remote tables or conditions, as appropriate to the verb. See the example and details.

Value

An H2OFrame expression (vector helpers return one column).

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      equal_sum(h_coalesce(as_h2o(x)[, "x"], 0), 16)
  }
}

Return the remote frame without downloading

Description

Row data remain in H2O. Base R handles expressions, column names, configuration and remote object handles. H2O may return schema metadata or scalar reductions to the client; no local table transformation is performed. This intentionally differs from the standalone prototype: it returns an H2OFrame, never a local table.

Usage

h_collect(.data)

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

Value

An H2OFrame expression (vector helpers return one column).

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      stopifnot(inherits(h_collect(x), "H2OFrame"))
  }
}

H2O count

Description

Usage

h_count(.data, ..., wt = NULL, sort = FALSE, name = "n", by = NULL)

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

...

Column selections, named expressions, remote tables or conditions, as appropriate to the verb. See the example and details.

wt

Optional weight column, captured as a name. Missing weights are removed.

sort

Order counts from largest to smallest.

name

Output count-column name.

by

Join column names; a named character vector maps left names to right names. For h_count, an optional grouping selection.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      z <- h_count(x, grp)
      rows(z, 3)
      equal_sum(as_h2o(z)[, "n"], 6)
  }
}

H2O cummax

Description

Row data remain in H2O. Base R handles expressions, column names, configuration and remote object handles. H2O may return schema metadata or scalar reductions to the client; no local table transformation is performed. Operations use current frame order. Arrange explicitly before windows. Grouped slicing and filling are rejected. Sampling is performed with H2O random columns; replacement uses a remote draw-index join.

Usage

h_cummax(x)

Arguments

x

An H2OFrame, an h2o_tbl, or (for selection helpers) a character pattern or vector. Local tables are not accepted.

Value

An H2OFrame expression (vector helpers return one column).

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      equal_sum(h_cummax(v), 210)
  }
}

H2O cummin

Description

Row data remain in H2O. Base R handles expressions, column names, configuration and remote object handles. H2O may return schema metadata or scalar reductions to the client; no local table transformation is performed. Operations use current frame order. Arrange explicitly before windows. Grouped slicing and filling are rejected. Sampling is performed with H2O random columns; replacement uses a remote draw-index join.

Usage

h_cummin(x)

Arguments

x

An H2OFrame, an h2o_tbl, or (for selection helpers) a character pattern or vector. Local tables are not accepted.

Value

An H2OFrame expression (vector helpers return one column).

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      equal_sum(h_cummin(v), 60)
  }
}

H2O cumsum

Description

Row data remain in H2O. Base R handles expressions, column names, configuration and remote object handles. H2O may return schema metadata or scalar reductions to the client; no local table transformation is performed. Operations use current frame order. Arrange explicitly before windows. Grouped slicing and filling are rejected. Sampling is performed with H2O random columns; replacement uses a remote draw-index join.

Usage

h_cumsum(x)

Arguments

x

An H2OFrame, an h2o_tbl, or (for selection helpers) a character pattern or vector. Local tables are not accepted.

Value

An H2OFrame expression (vector helpers return one column).

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      equal_sum(h_cumsum(v), 560)
  }
}

Deduplicate remote rows

Description

Usage

h_distinct(.data, ..., .keep_all = FALSE)

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

...

Column selections, named expressions, remote tables or conditions, as appropriate to the verb. See the example and details.

.keep_all

Keep every column, retaining the first row for each distinct key.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      rows(h_distinct(x, grp), 3)
  }
}

H2O drop na

Description

Usage

h_drop_na(.data, ...)

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

...

Column selections, named expressions, remote tables or conditions, as appropriate to the verb. See the example and details.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      rows(h_drop_na(p), 333)
  }
}

H2O explain

Description

Row data remain in H2O. Base R handles expressions, column names, configuration and remote object handles. H2O may return schema metadata or scalar reductions to the client; no local table transformation is performed. Displays dimensions, column metadata and grouping only. No row preview or cost estimate is downloaded.

Usage

h_explain(.data)

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      stopifnot(identical(h_explain(x), x))
  }
}

Fill missing values in row order

Description

Row data remain in H2O. Base R handles expressions, column names, configuration and remote object handles. H2O may return schema metadata or scalar reductions to the client; no local table transformation is performed. Operations use current frame order. Arrange explicitly before windows. Grouped slicing and filling are rejected. Sampling is performed with H2O random columns; replacement uses a remote draw-index join.

Usage

h_fill(.data, ..., .direction = c("down", "up"))

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

...

Column selections, named expressions, remote tables or conditions, as appropriate to the verb. See the example and details.

.direction

Fill down from the preceding value, or up from the following value.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      equal_sum(as_h2o(h_fill(x, all_of("x")))[, "x"], 20)
  }
}

Keep rows satisfying remote predicates

Description

Usage

h_filter(.data, ...)

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

...

Column selections, named expressions, remote tables or conditions, as appropriate to the verb. See the example and details.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      rows(h_filter(x, x >= 3), 3)
  }
}

Exclude rows satisfying remote predicates

Description

Usage

h_filter_out(.data, ...)

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

...

Column selections, named expressions, remote tables or conditions, as appropriate to the verb. See the example and details.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      rows(h_filter_out(x, x >= 3), 2)
  }
}

H2O full join

Description

Row data remain in H2O. Base R handles expressions, column names, configuration and remote object handles. H2O may return schema metadata or scalar reductions to the client; no local table transformation is performed. Equijoins use H2O merge. Missing-key behavior follows H2O, not SQL NULL semantics. Output row order is unspecified; sort explicitly. Full joins combine a left join and unmatched right rows. Character keys are converted to categorical keys remotely. Suffix collisions and violated relationship constraints raise errors.

Usage

h_full_join(x, y, by = NULL, suffix = c(".x", ".y"), relationship = NULL)

Arguments

x

An H2OFrame, an h2o_tbl, or (for selection helpers) a character pattern or vector. Local tables are not accepted.

y

A second remote table for joins, or a scalar comparison value for h_na_if.

by

Join column names; a named character vector maps left names to right names. For h_count, an optional grouping selection.

suffix

Two suffixes for overlapping non-key column names.

relationship

Optional one-to-one, one-to-many, many-to-one or many-to-many key constraint.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      rows(h_full_join(h_slice_head(x, 3), h_slice_tail(x, 4), by = "id"), 6)
  }
}

H2O glimpse

Description

Row data remain in H2O. Base R handles expressions, column names, configuration and remote object handles. H2O may return schema metadata or scalar reductions to the client; no local table transformation is performed. Displays dimensions, column metadata and grouping only. No row preview or cost estimate is downloaded.

Usage

h_glimpse(.data, width = 60)

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

width

Optional zero-padding width for num_range; retained metadata-display width for h_glimpse.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      stopifnot(identical(h_glimpse(x), x))
  }
}

H2O group by

Description

Usage

h_group_by(.data, ..., .add = FALSE)

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

...

Column selections, named expressions, remote tables or conditions, as appropriate to the verb. See the example and details.

.add

Add columns to existing grouping metadata rather than replace it.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      stopifnot(identical(h_group_by(x, grp)$groups, "grp"))
  }
}

H2O group keys

Description

Usage

h_group_keys(.data)

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      rows(h_group_keys(h_group_by(x, grp)), 3)
  }
}

H2O group vars

Description

Usage

h_group_vars(.data)

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

Value

A character vector of grouping column names.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      stopifnot(identical(h_group_vars(h_group_by(x, grp)), "grp"))
  }
}

H2O if else

Description

Usage

h_if_else(condition, true_val, false_val, missing = NA)

Arguments

condition

A one-column H2OFrame of conditions.

true_val

Scalar or compatible H2O column for true conditions.

false_val

Scalar or compatible H2O column for false conditions.

missing

Value for missing conditions.

Value

An H2OFrame expression (vector helpers return one column).

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      equal_sum(h_if_else(v >= 40, 1, 0), 3)
  }
}

H2O inner join

Description

Row data remain in H2O. Base R handles expressions, column names, configuration and remote object handles. H2O may return schema metadata or scalar reductions to the client; no local table transformation is performed. Equijoins use H2O merge. Missing-key behavior follows H2O, not SQL NULL semantics. Output row order is unspecified; sort explicitly. Full joins combine a left join and unmatched right rows. Character keys are converted to categorical keys remotely. Suffix collisions and violated relationship constraints raise errors.

Usage

h_inner_join(x, y, by = NULL, suffix = c(".x", ".y"), relationship = NULL)

Arguments

x

An H2OFrame, an h2o_tbl, or (for selection helpers) a character pattern or vector. Local tables are not accepted.

y

A second remote table for joins, or a scalar comparison value for h_na_if.

by

Join column names; a named character vector maps left names to right names. For h_count, an optional grouping selection.

suffix

Two suffixes for overlapping non-key column names.

relationship

Optional one-to-one, one-to-many, many-to-one or many-to-many key constraint.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      rows(h_inner_join(h_slice_head(x, 3), h_slice_tail(x, 4), by = "id"), 1)
  }
}

H2O lag

Description

Row data remain in H2O. Base R handles expressions, column names, configuration and remote object handles. H2O may return schema metadata or scalar reductions to the client; no local table transformation is performed. Operations use current frame order. Arrange explicitly before windows. Grouped slicing and filling are rejected. Sampling is performed with H2O random columns; replacement uses a remote draw-index join.

Usage

h_lag(x, n = 1, default = NA)

Arguments

x

An H2OFrame, an h2o_tbl, or (for selection helpers) a character pattern or vector. Local tables are not accepted.

n

A nonnegative integer row count or shift distance.

default

Padding value for lag or lead.

Value

An H2OFrame expression (vector helpers return one column).

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      equal_sum(h_lag(v, default = 0), 150)
  }
}

H2O lead

Description

Row data remain in H2O. Base R handles expressions, column names, configuration and remote object handles. H2O may return schema metadata or scalar reductions to the client; no local table transformation is performed. Operations use current frame order. Arrange explicitly before windows. Grouped slicing and filling are rejected. Sampling is performed with H2O random columns; replacement uses a remote draw-index join.

Usage

h_lead(x, n = 1, default = NA)

Arguments

x

An H2OFrame, an h2o_tbl, or (for selection helpers) a character pattern or vector. Local tables are not accepted.

n

A nonnegative integer row count or shift distance.

default

Padding value for lag or lead.

Value

An H2OFrame expression (vector helpers return one column).

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      equal_sum(h_lead(v, default = 0), 200)
  }
}

H2O left join

Description

Row data remain in H2O. Base R handles expressions, column names, configuration and remote object handles. H2O may return schema metadata or scalar reductions to the client; no local table transformation is performed. Equijoins use H2O merge. Missing-key behavior follows H2O, not SQL NULL semantics. Output row order is unspecified; sort explicitly. Full joins combine a left join and unmatched right rows. Character keys are converted to categorical keys remotely. Suffix collisions and violated relationship constraints raise errors.

Usage

h_left_join(x, y, by = NULL, suffix = c(".x", ".y"), relationship = NULL)

Arguments

x

An H2OFrame, an h2o_tbl, or (for selection helpers) a character pattern or vector. Local tables are not accepted.

y

A second remote table for joins, or a scalar comparison value for h_na_if.

by

Join column names; a named character vector maps left names to right names. For h_count, an optional grouping selection.

suffix

Two suffixes for overlapping non-key column names.

relationship

Optional one-to-one, one-to-many, many-to-one or many-to-many key constraint.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      rows(h_left_join(h_slice_head(x, 3), h_slice_tail(x, 4), by = "id"), 3)
  }
}

Create or replace columns remotely

Description

Usage

h_mutate(.data, ...)

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

...

Column selections, named expressions, remote tables or conditions, as appropriate to the verb. See the example and details.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      z <- h_mutate(x, double = y * 2, next_value = double + 1)
      equal_sum(as_h2o(z)[, "next_value"], 426)
  }
}

H2O mutate across

Description

Usage

h_mutate_across(.data, .cols, .fns, .names = "{col}_{fn}")

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

.cols

A column selection or character vector of column names.

.fns

A function or named list. Summary functions must be supported aggregates; mutation functions must have one argument and a supported expression body.

.names

Output naming template using col and fn tokens enclosed in braces.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      z <- h_mutate_across(x, c("x1", "x2"), list(double = function(z) z * 2))
      equal_sum(as_h2o(z)[, "x2_double"], 1140)
  }
}

H2O na if

Description

Usage

h_na_if(x, y)

Arguments

x

An H2OFrame, an h2o_tbl, or (for selection helpers) a character pattern or vector. Local tables are not accepted.

y

A second remote table for joins, or a scalar comparison value for h_na_if.

Value

An H2OFrame expression (vector helpers return one column).

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      stopifnot(h2o::h2o.sum(is.na(h_na_if(v, 10))) == 1)
  }
}

Create a list of remote group handles

Description

Row data remain in H2O. Base R handles expressions, column names, configuration and remote object handles. H2O may return schema metadata or scalar reductions to the client; no local table transformation is performed. Returns a list with keys (h2o_tbl) and data (list of h2o_tbl handles), not a data frame with a list-column. This makes one remote join per group and is intended for modest group counts.

Usage

h_nest_by(.data, ...)

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

...

Column selections, named expressions, remote tables or conditions, as appropriate to the verb. See the example and details.

Value

A list containing remote keys and a list of remote table handles.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      z <- h_nest_by(x, grp)
      stopifnot(length(z$data) == 3)
      invisible(lapply(z$data, rows, n = 2))
  }
}

Upload the bundled Palmer penguins CSV

Description

Usage

h_penguins()

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      rows(p, 344)
  }
}

H2O pick

Description

Usage

h_pick(.data, ...)

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

...

Column selections, named expressions, remote tables or conditions, as appropriate to the verb. See the example and details.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      cols(h_pick(x, starts_with("x")), c("x", "x1", "x2"))
  }
}

H2O pivot longer

Description

Usage

h_pivot_longer(.data, cols, names_to = "name", values_to = "value")

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

cols

Columns to reshape from wide to long.

names_to

Name of the output column containing former column names.

values_to

Name of the output value column.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      rows(h_pivot_longer(h_select(x, id, x1, x2), c("x1", "x2")), 12)
  }
}

Widen unique ID and name combinations

Description

Row data remain in H2O. Base R handles expressions, column names, configuration and remote object handles. H2O may return schema metadata or scalar reductions to the client; no local table transformation is performed. Duplicate ID/name combinations and missing names are rejected; summarize first. Category levels are schema metadata. All output cells are formed by remote joins. values_fill replaces both absent combinations and existing missing cells. No list-columns or implicit aggregation are supported.

Usage

h_pivot_wider(.data, names_from, values_from, id_cols = NULL, values_fill = NULL)

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

names_from

One column whose category domain becomes output column names.

values_from

One column containing cell values.

id_cols

ID column selection; defaults to all columns other than names_from and values_from.

values_fill

Optional replacement for missing output cells.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      long <- h_pivot_longer(h_select(x, id, x1, x2), c("x1", "x2"))
      wide <- h_pivot_wider(long, name, value)
      rows(wide, 6)
      equal_sum(as_h2o(wide)[, "x2"], 570)
  }
}

H2O relocate

Description

Usage

h_relocate(.data, ..., .before = NULL, .after = NULL)

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

...

Column selections, named expressions, remote tables or conditions, as appropriate to the verb. See the example and details.

.before

Column selection before which the selected columns are moved.

.after

Column selection after which the selected columns are moved.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      stopifnot(names(as_h2o(h_relocate(x, y, .before = id)))[1] == "y")
  }
}

H2O rename

Description

Usage

h_rename(.data, ...)

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

...

Column selections, named expressions, remote tables or conditions, as appropriate to the verb. See the example and details.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      cols(h_rename(h_select(x, id), identifier = id), "identifier")
  }
}

H2O replace na

Description

Usage

h_replace_na(x, replace)

Arguments

x

An H2OFrame, an h2o_tbl, or (for selection helpers) a character pattern or vector. Local tables are not accepted.

replace

For h_replace_na, a scalar or named list of replacements. For h_slice_sample, whether to sample with replacement.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      equal_sum(as_h2o(h_replace_na(x, list(x = 5)))[, "x"], 21)
  }
}

H2O replace na cols

Description

Usage

h_replace_na_cols(.data, replace, ...)

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

replace

For h_replace_na, a scalar or named list of replacements. For h_slice_sample, whether to sample with replacement.

...

Column selections, named expressions, remote tables or conditions, as appropriate to the verb. See the example and details.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      equal_sum(as_h2o(h_replace_na_cols(x, 5, all_of("x")))[, "x"], 21)
  }
}

H2O right join

Description

Row data remain in H2O. Base R handles expressions, column names, configuration and remote object handles. H2O may return schema metadata or scalar reductions to the client; no local table transformation is performed. Equijoins use H2O merge. Missing-key behavior follows H2O, not SQL NULL semantics. Output row order is unspecified; sort explicitly. Full joins combine a left join and unmatched right rows. Character keys are converted to categorical keys remotely. Suffix collisions and violated relationship constraints raise errors.

Usage

h_right_join(x, y, by = NULL, suffix = c(".x", ".y"), relationship = NULL)

Arguments

x

An H2OFrame, an h2o_tbl, or (for selection helpers) a character pattern or vector. Local tables are not accepted.

y

A second remote table for joins, or a scalar comparison value for h_na_if.

by

Join column names; a named character vector maps left names to right names. For h_count, an optional grouping selection.

suffix

Two suffixes for overlapping non-key column names.

relationship

Optional one-to-one, one-to-many, many-to-one or many-to-many key constraint.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      rows(h_right_join(h_slice_head(x, 3), h_slice_tail(x, 4), by = "id"), 4)
  }
}

Select columns using schema metadata

Description

Usage

h_select(.data, ...)

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

...

Column selections, named expressions, remote tables or conditions, as appropriate to the verb. See the example and details.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      cols(h_select(x, id, grp), c("id", "grp"))
  }
}

H2O semi join

Description

Row data remain in H2O. Base R handles expressions, column names, configuration and remote object handles. H2O may return schema metadata or scalar reductions to the client; no local table transformation is performed. Equijoins use H2O merge. Missing-key behavior follows H2O, not SQL NULL semantics. Output row order is unspecified; sort explicitly. Full joins combine a left join and unmatched right rows. Character keys are converted to categorical keys remotely. Suffix collisions and violated relationship constraints raise errors.

Usage

h_semi_join(x, y, by = NULL)

Arguments

x

An H2OFrame, an h2o_tbl, or (for selection helpers) a character pattern or vector. Local tables are not accepted.

y

A second remote table for joins, or a scalar comparison value for h_na_if.

by

Join column names; a named character vector maps left names to right names. For h_count, an optional grouping selection.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      rows(h_semi_join(h_slice_head(x, 3), h_slice_tail(x, 4), by = "id"), 1)
  }
}

H2O separate

Description

Row data remain in H2O. Base R handles expressions, column names, configuration and remote object handles. H2O may return schema metadata or scalar reductions to the client; no local table transformation is performed. H2O splits using a regular expression. The maximum number of split fields must equal length(into); shorter rows are padded by H2O. The source column is removed.

Usage

h_separate(.data, col, into, sep = "-")

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

col

The source column for h_separate, or the new output name for h_unite.

into

Names of split output columns.

sep

Regular expression separator for h_separate; literal separator for h_unite (see its restrictions).

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      z <- h_separate(x, txt, c("letter", "number"))
      rows(z, 6)
      stopifnot(all(c("letter", "number") %in% names(as_h2o(z))))
  }
}

H2O slice head

Description

Row data remain in H2O. Base R handles expressions, column names, configuration and remote object handles. H2O may return schema metadata or scalar reductions to the client; no local table transformation is performed. Operations use current frame order. Arrange explicitly before windows. Grouped slicing and filling are rejected. Sampling is performed with H2O random columns; replacement uses a remote draw-index join.

Usage

h_slice_head(.data, n = 6)

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

n

A nonnegative integer row count or shift distance.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      rows(h_slice_head(x, 2), 2)
  }
}

H2O slice max

Description

Row data remain in H2O. Base R handles expressions, column names, configuration and remote object handles. H2O may return schema metadata or scalar reductions to the client; no local table transformation is performed. Operations use current frame order. Arrange explicitly before windows. Grouped slicing and filling are rejected. Sampling is performed with H2O random columns; replacement uses a remote draw-index join.

Usage

h_slice_max(.data, order_by, n = 1, with_ties = TRUE)

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

order_by

One unquoted ordering column.

n

A nonnegative integer row count or shift distance.

with_ties

Include all rows tied at the selected cutoff. Missing ordering values are excluded.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      equal_sum(as_h2o(h_slice_max(x, y, n = 2))[, "y"], 110)
  }
}

H2O slice min

Description

Row data remain in H2O. Base R handles expressions, column names, configuration and remote object handles. H2O may return schema metadata or scalar reductions to the client; no local table transformation is performed. Operations use current frame order. Arrange explicitly before windows. Grouped slicing and filling are rejected. Sampling is performed with H2O random columns; replacement uses a remote draw-index join.

Usage

h_slice_min(.data, order_by, n = 1, with_ties = TRUE)

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

order_by

One unquoted ordering column.

n

A nonnegative integer row count or shift distance.

with_ties

Include all rows tied at the selected cutoff. Missing ordering values are excluded.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      equal_sum(as_h2o(h_slice_min(x, y, n = 2))[, "y"], 30)
  }
}

H2O slice sample

Description

Row data remain in H2O. Base R handles expressions, column names, configuration and remote object handles. H2O may return schema metadata or scalar reductions to the client; no local table transformation is performed. Operations use current frame order. Arrange explicitly before windows. Grouped slicing and filling are rejected. Sampling is performed with H2O random columns; replacement uses a remote draw-index join.

Usage

h_slice_sample(.data, n = NULL, prop = NULL, replace = FALSE, seed = NULL)

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

n

A nonnegative integer row count or shift distance.

prop

Sample fraction. Supply either n or prop.

replace

For h_replace_na, a scalar or named list of replacements. For h_slice_sample, whether to sample with replacement.

seed

Optional H2O random seed. With NULL (the default), h2o.runif() is called without a seed argument so H2O chooses the seed. A supplied seed is forwarded unchanged; reproducibility assumes the same cluster layout and input frame.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      rows(h_slice_sample(x, n = 3, seed = 42), 3)
      rows(h_slice_sample(x, n = 9, replace = TRUE, seed = 42), 9)
  }
}

H2O slice tail

Description

Row data remain in H2O. Base R handles expressions, column names, configuration and remote object handles. H2O may return schema metadata or scalar reductions to the client; no local table transformation is performed. Operations use current frame order. Arrange explicitly before windows. Grouped slicing and filling are rejected. Sampling is performed with H2O random columns; replacement uses a remote draw-index join.

Usage

h_slice_tail(.data, n = 6)

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

n

A nonnegative integer row count or shift distance.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      equal_sum(as_h2o(h_slice_tail(x, 2))[, "y"], 110)
  }
}

Compute grouped or global H2O summaries

Description

Row data remain in H2O. Base R handles expressions, column names, configuration and remote object handles. H2O may return schema metadata or scalar reductions to the client; no local table transformation is performed. Supported reductions: mean, sum, min, max, sd, var, median, n() and n_distinct(). A reduction accepts one column and optional na.rm. Summary results are ungrouped. Empty-input summaries raise an error. Across summaries use the default na.rm = FALSE; drop missing rows explicitly when needed.

Usage

h_summarise(.data, ..., .by = NULL)

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

...

Column selections, named expressions, remote tables or conditions, as appropriate to the verb. See the example and details.

.by

Per-call grouping selection on ungrouped input. Grouping does not persist in the result.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      z <- h_summarise(x, n = n())
      equal_sum(as_h2o(z), 6)
  }
}

Compute grouped or global H2O summaries

Description

Row data remain in H2O. Base R handles expressions, column names, configuration and remote object handles. H2O may return schema metadata or scalar reductions to the client; no local table transformation is performed. Supported reductions: mean, sum, min, max, sd, var, median, n() and n_distinct(). A reduction accepts one column and optional na.rm. Summary results are ungrouped. Empty-input summaries raise an error. Across summaries use the default na.rm = FALSE; drop missing rows explicitly when needed.

Usage

h_summarize(.data, ..., .by = NULL)

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

...

Column selections, named expressions, remote tables or conditions, as appropriate to the verb. See the example and details.

.by

Per-call grouping selection on ungrouped input. Grouping does not persist in the result.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      z <- h_summarize(x, n = n(), avg = mean(y), .by = grp)
      rows(z, 3)
      equal_sum(as_h2o(z)[, "avg"], 105)
  }
}

H2O summarize across

Description

Row data remain in H2O. Base R handles expressions, column names, configuration and remote object handles. H2O may return schema metadata or scalar reductions to the client; no local table transformation is performed. Supported reductions: mean, sum, min, max, sd, var, median, n() and n_distinct(). A reduction accepts one column and optional na.rm. Summary results are ungrouped. Empty-input summaries raise an error. Across summaries use the default na.rm = FALSE; drop missing rows explicitly when needed.

Usage

h_summarize_across(.data, .cols, .fns, .names = "{col}_{fn}")

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

.cols

A column selection or character vector of column names.

.fns

A function or named list. Summary functions must be supported aggregates; mutation functions must have one argument and a supported expression body.

.names

Output naming template using col and fn tokens enclosed in braces.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      z <- h_summarize_across(h_group_by(x, grp), "y", list(mean = mean))
      equal_sum(as_h2o(z)[, "y_mean"], 105)
  }
}

H2O transmute

Description

Usage

h_transmute(.data, ...)

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

...

Column selections, named expressions, remote tables or conditions, as appropriate to the verb. See the example and details.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      z <- h_transmute(x, kg = y/1000)
      cols(z, "kg")
      equal_sum(as_h2o(z), 0.21)
  }
}

H2O ungroup

Description

Usage

h_ungroup(.data)

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      stopifnot(length(h_group_vars(h_ungroup(h_group_by(x, grp)))) == 0)
  }
}

Concatenate categorical values on H2O

Description

Row data remain in H2O. Base R handles expressions, column names, configuration and remote object handles. H2O may return schema metadata or scalar reductions to the client; no local table transformation is performed. Uses H2O interaction with underscore encoding. Missing inputs produce missing output. A reserved encoding token is rejected. Separators containing dollar signs or backslashes are rejected because H2O uses Java replacement syntax. Interaction domains can be expensive for high-cardinality inputs.

Usage

h_unite(.data, col, ..., sep = "_")

Arguments

.data

An h2o_tbl. Its frame remains on the H2O cluster.

col

The source column for h_separate, or the new output name for h_unite.

...

Column selections, named expressions, remote tables or conditions, as appropriate to the verb. See the example and details.

sep

Regular expression separator for h_separate; literal separator for h_unite (see its restrictions).

Value

An h2o_tbl holding a remote H2OFrame; inspection helpers return it invisibly.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      z <- h_unite(h_separate(x, txt, c("letter", "number")), "text", letter, number, 
          sep = "-")
      rows(h_filter(z, text == "b-2"), 1)
  }
}

H2O last col

Description

Usage

last_col(offset = 0)

Arguments

offset

A nonnegative offset from the last column.

Value

A column-selection specification containing metadata only.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      cols(h_select(x, last_col()), "x2")
  }
}

H2O matches

Description

Usage

matches(x)

Arguments

x

An H2OFrame, an h2o_tbl, or (for selection helpers) a character pattern or vector. Local tables are not accepted.

Value

A column-selection specification containing metadata only.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      cols(h_select(x, matches("^x[12]$")), c("x1", "x2"))
  }
}

H2O num range

Description

Usage

num_range(prefix, range, width = NULL)

Arguments

prefix

A column-name prefix.

range

Numeric suffixes to match.

width

Optional zero-padding width for num_range; retained metadata-display width for h_glimpse.

Value

A column-selection specification containing metadata only.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      cols(h_select(x, num_range("x", 1:2)), c("x1", "x2"))
  }
}

H2O starts with

Description

Usage

starts_with(x)

Arguments

x

An H2OFrame, an h2o_tbl, or (for selection helpers) a character pattern or vector. Local tables are not accepted.

Value

A column-selection specification containing metadata only.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      cols(h_select(x, starts_with("x")), c("x", "x1", "x2"))
  }
}

H2O when all

Description

Usage

when_all(...)

Arguments

...

Column selections, named expressions, remote tables or conditions, as appropriate to the verb. See the example and details.

Value

An H2OFrame expression (vector helpers return one column).

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      rows(h_filter(x, when_all(y >= 20, y <= 40)), 3)
  }
}

H2O when any

Description

Usage

when_any(...)

Arguments

...

Column selections, named expressions, remote tables or conditions, as appropriate to the verb. See the example and details.

Value

An H2OFrame expression (vector helpers return one column).

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      rows(h_filter(x, when_any(y == 10, y == 60)), 2)
  }
}

H2O where

Description

Usage

where(predicate)

Arguments

predicate

One of base is.numeric, is.character or is.factor; inspected through H2O schema metadata.

Value

A column-selection specification containing metadata only.

Examples

# Connect explicitly before running these examples:
# h2o::h2o.init(nthreads = 2, max_mem_size = "2G")
if (h2o_available()) {
  x <- h2o_tbl(h2o::h2o.uploadFile(
    system.file("extdata", "toy.csv", package = "h2oplyr"),
    col.types = c("numeric", "enum", "numeric", "numeric",
                  "string", "numeric", "numeric")))
  p <- h_penguins()
  v <- as_h2o(x)[, "y"]
  rows <- function(z, n) stopifnot(nrow(as_h2o(z)) == n)
  cols <- function(z, nm) stopifnot(identical(names(as_h2o(z)), nm))
  equal_sum <- function(z, value) stopifnot(isTRUE(all.equal(
    h2o::h2o.sum(z, na.rm = TRUE), value, check.attributes = FALSE)))
  {
      cols(h_select(x, where(is.numeric)), c("id", "x", "y", "x1", "x2"))
  }
}

Package {h2oplyr}

Remote H2O Data Manipulation

Description

See Also

H2O %.%

Description

Usage

Arguments

Value

Examples

H2O all of

Description

Usage

Arguments

Value

Examples

H2O any of

Description

Usage

Arguments

Value

Examples

Access the remote H2O frame

Description

Usage

Arguments

Value

Examples

H2O contains

Description

Usage

Arguments

Value

Examples

H2O desc

Description

Usage

Arguments

Value

Examples

H2O ends with

Description

Usage

Arguments

Value

Examples

H2O everything

Description

Usage

Value

Examples

Check an existing H2O connection

Description

Usage

Value

Examples

Wrap an existing remote H2O frame

Description

Usage

Arguments

Value

Examples

H2O across

Description

Usage

Arguments

Value

Examples

H2O anti join

Description

Usage

Arguments

Value

Examples

H2O arrange

Description

Usage

Arguments

Value

Examples