transform_to_transactional: transform_to_transactional function

In yForecasting/salesforecasting: Sales Forecasting

transform_to_transactional function

Description

Converts a data frame with a ID column and all different date columns to a transactional data frame with three columns: ID, Date and Value. The function removes zero_demand observations to minimize storage size.

Usage

transform_to_transactional(
  df,
  keep_columns = TRUE,
  non_zero_demand = FALSE,
  start = NULL,
  end = NULL,
  freq = NULL
)

Arguments

df	A dataframe or matrix containing sales on a particular date. Columns stand for time observations
keep_columns	A boolean stating if the function returns the original columns. Default is TRUE.
non_zero_demand	A boolean stating if the function should returns zero observations. Default is FALSE.
start	A date object indicating the starting date in the transactional df. NOTE 1: Only considered if keep_columns is set to FALSE. NOTE 2: If keep_columns is set to false, the start parameter should be given along with end or freq
end	A date object indicating the last date on the transcational df
freq	The frequency of the observated observations. Can be numeric, or string (eg 'day', 'week', '2 weeks', 'month', '2 months', 'quater', 'year')

Value

A data frame or matrix with three columns: ID, Date and Value

Author(s)

Filotas Theodosiou

Examples

dates <- c('2019-07-01', '2019-07-02', '2019-07-03', '2019-07-01')
stores <- c('AM','AC')
example_df <- data.frame( t1 = c(15, 12),
                 t2 = c(30, 40 ),
                 t3 = c(0, 5),
                 t4 = c(21, 30)
                 )
rownames(example_df) <- stores
tr_df <- transform_to_transactional(example_df, keep_columns = TRUE) # Keeps current columns
tr_df_dates <- transform_to_transactional(example_df, keep_columns = FALSE,
                                          start = as.Date(dates[1]), freq = 'day' )

yForecasting/salesforecasting documentation built on April 29, 2022, 7:21 p.m.