maditr (version 0.7.4)

to_list: Apply an expression to each element of a list or vector

Description

• to_list always returns a list, each element of which is the result of expression expr on the elements of data. By default, NULL's will be removed from the result. You can change this behavior with skip_null argument.

• to_vec is the same as to_list but tries to convert its result to vector via unlist.

• to_df and to_dfr try to combine its results to data.table by rows.

• to_dfc tries to combine its result to data.table by columns.

Expression can use predefined variables: '.x' is a value of current list element, '.name' is a name of the element and '.index' is sequential number of the element. '.value' is an alias to '.x'.

Usage

to_list(
  data,
  expr = NULL,
  ...,
  skip_null = TRUE,
  trace = FALSE,
  trace_step = 1L
)
to_vec(
  data,
  expr = NULL,
  ...,
  skip_null = TRUE,
  trace = FALSE,
  trace_step = 1L,
  recursive = TRUE,
  use.names = TRUE
)
to_df(
  data,
  expr = NULL,
  ...,
  trace = FALSE,
  trace_step = 1L,
  idvalue = NULL,
  idname = "item_id"
)
to_dfr(
  data,
  expr = NULL,
  ...,
  trace = FALSE,
  trace_step = 1L,
  idvalue = NULL,
  idname = "item_id"
)
to_dfc(data, expr = NULL, ..., trace = FALSE, trace_step = 1)

Arguments

data

data.frame/list/vector

expr

expression or function. Expression can use predefined variables: '.x' is a value of current list element, '.name' is a name of the element and '.index' is sequential number of the element.

...

further arguments provided if 'expr' is function.

skip_null

logical Should we skip NULL's from result? Default is TRUE

trace

FALSE by default. Should we report progress during execution? Possible values are TRUE, FALSE, "pb" (progress bar) or custom expression in 'quote', e. g. 'quote(print(.x))'. Expression can contain '.x', '.name', '.res' and '.index' variables.

trace_step

integer. 1 by default. Step for reporting progress. Ignored if 'trace' argument is equal to FALSE.

recursive

logical. Should unlisting be applied to list components of x? For details see unlist.

use.names

logical. TRUE by default. Should names of source list be preserved? Setting it to FALSE in some cases can greatly increase performance. For details see unlist.

idvalue

expression for calculation id column. Usually it is just unquoted symbols: one of the '.name', '.index' or '.x'.

idname

character, 'item_id' by default. Name for the id column.

Value

'to_list' returns list, 'to_vec' tries to return vector and other functions return data.table

Examples

# NOT RUN {
1:5 %>%
    to_list(rnorm(n = 3, .x))

# or in 'lapply' style
1:5 %>%
    to_list(rnorm, n = 3) %>%
    to_vec(mean)

# or use an anonymous function
1:5 %>%
    to_list(function(x) rnorm(3, x))

# Use to_vec() to reduce output to a vector instead
# of a list:
# filtering - return only even numbers
to_vec(1:10, if(.x %% 2 == 0) .x)

# filtering - mean only on the numeric columns
to_vec(iris, if(is.numeric(.x)) mean(.x))

# mean for numerics, number of distincts for others
to_vec(iris, if(is.numeric(.x)) mean(.x) else uniqueN(.x))

# means for Sepal
to_vec(iris, if(startsWith(.name, "Sepal")) mean(.x))

# A more realistic example: split a data frame into pieces, fit a
# model to each piece, summarise and extract R^2
mtcars %>%
    split(.$cyl) %>%
    to_list(summary(lm(mpg ~ wt, data = .x))) %>%
    to_vec(.x$r.squared)

# If each element of the output is a data frame, use
# to_df to row-bind them together:
mtcars %>%
    split(.$cyl) %>%
    to_list(lm(mpg ~ wt, data = .x)) %>%
    to_df(c(cyl = .name, coef(.x)))

# }
# NOT RUN {
# read all csv files in "data" to data.frame
all_files = dir("data", pattern = "csv$", full.names = TRUE) %>%
    to_df(fread,
          idvalue = basename(.x),
          idname = "filename",
          trace = "pb"
          )
# }