pipe_function: Wrapper for putting a single function into a pipeline

In jeroenvdhoven/datapiper: datapiper

Description

Wrapper for putting a single function into a pipeline

Usage

pipe_function(train, f, ...)

Arguments

train	Data frame containing the train data.
f	The function to be put into the pipeline. It is important that the function can be applied to new datasets without using any information from the train dataset, e.g. lowercasing column names. It should take a data argumet
...	Additional arguments to be provided to f

Value

A list of the transformed train dataset and a .predict function to be used on new data.

Examples

data <- dplyr::data_frame(var = 0, Var = 0, camelCase = 0, good_name = 0,
                          `0none.` = 0, `bad  ` = 0, `j&d` = 0, `spac ed` = 0)
pipe_function(data, standard_column_names)

# You can also use this to append a custom model to the pipeline
data <- dplyr::data_frame(x = 1:10, y = (1:10) + rnorm(10))
model <- lm(y ~ x, data)

self_contained_function <- function(data) predict(model, data)
model_pipe <- pipe_function(data, self_contained_function)

predictions <- invoke(model_pipe$pipe, data)

jeroenvdhoven/datapiper documentation built on July 14, 2019, 9:34 p.m.