long: Create a long file from a wide file
In gmonette/spidanew: New Functions for the spida Package

Description Usage Arguments Details Value Examples

Uses a minimal number of arguments to create a long file using stats::reshape. Produces output even when long variable names and time values are not fully crossed.

1 2	long(data, sep = "_", timevar = "time", idvar = "id", ids = 1:nrow(data), expand = TRUE, safe_sep = "#%@!", ...)

`data`	wide data frame
`sep`	(default '_') single character separator between long names and 'time' value. Variable names with this separator are transformed to long variables.
`timevar`	(default 'time') names of variable in output long file to identify occasions. Its values are taken from the suffix following the 'sep' character in each time-varying variable.
`idvar`	(default: 'id') variable name used in output long to identify row in input wide file. It may exist in the input wide file and must, in that case, have a unique value in each row. If not, its values are replaced with row numbers.
`ids`	(default `1:nrow(data)`) values for idvar in long file if the variable `idvar` does not exist in in the input wide file. Ignored if `idvar` exists in `data`.
`expand`	(default TRUE): if 'time' values are inconsistent, fill in missing 'time's with NAs.
`safe_sep`	temporary safe? separator
`...`	additional parameters are passed to `reshape`.

long is intended for the simple case in which 'wide' variables in the input data frame are identified by the fact that they contain a separator character that separates the name of the variable in the long file from the value of the 'time' variable that identifies the corresponding row in the long file, e.g x_1, x_2, x_3 or brain.volume_left, brain.volume_right. If the separator ('_' by default) occurs in other variables, it must be temporarily substituted.

rehape does not work if long variable names and time values are not fully crossed, e.g x_1, x_2, x_3, y_1, y_2. By default long creates additional variables with "NAs" so the set of variables given to reshape is fully crossed, e.g. adding a variable y_3 <- NA.

'long' file with each wide row repeated as many times as there are distinct values for the 'timevar' variable.

z <- data.frame(id =letters[1:10], id2= 11:20, v_L = 1:10, v_R = 11:20)
long(z)
long(z, timevar = 'Side', idvar = 'idn', ids = LETTERS[1:10])
long(z, timevar = 'Side', idvar = 'idn', ids = z$id2)

# unbalanced times
z <- data.frame(id =letters[1:10], id2= 11:20, v_L = 1:10, v_R = 11:20, z_L = 21:30)
long(z)