replyr (version 0.2.0)

replyr_spread: Spread values found in rowControlColumn row groups as new columns. A concept demonstration, not a usable function.

Description

Spread values found in rowControlColumn row groups as new columns. Values types (new column names) are identified in measurementNameColumn and values are taken from measurementValueColumn. This is denormalizing operation, or essentially a tidyr::spread, dplyr::dcast, or pivot. This implementation moves so much data it is essentially working locally and also very inefficient.

Usage

replyr_spread(df, rowControlColumn, measurementNameColumn, measurementValueColumn, maxcols = 100, useTidyr = FALSE, eagerCompute = FALSE)

Arguments

df

data item

rowControlColumn

column to determine which sets of rows are considered a group.

measurementNameColumn

column to take measurement names from (values become new columns)

measurementValueColumn

column to take measurement values from

maxcols

maximum number of values to expand to columns

useTidyr

if TRUE use tidyr instead of calculating on own (only works on local data types)

eagerCompute

if TRUE call compute on intermediate results

Value

data item

Examples

d <- data.frame(
  index = c(1, 2, 3, 1, 2, 3),
  meastype = c('meas1','meas1','meas1','meas2','meas2','meas2'),
  meas = c('m1_1', 'm1_2', 'm1_3', 'm2_1', 'm2_2', 'm2_3'),
  stringsAsFactors = FALSE)
replyr_spread(d,'index','meastype','meas')
replyr_spread(d,'index','meastype','meas',useTidyr=TRUE)