quoted csv

[certara-mtomashevskiy]

if someone exports rds presented like this
write.csv(rds01, 'rds01.csv')
and then tries to import it again
method.A(path.in = getwd(), file = 'rds01', ext = 'csv')
an error is given that column names are not found:
Error in get.data(path.in = path.in, path.out = path.out, file = file, :
Column names must be given as 'subject', 'period', 'sequence', 'treatment'.
since in fact the column names are quoted
To avoid this the user should use
write.csv(rds01, 'rds01.csv', quote = FALSE)

[certara-mtomashevskiy]

We can avoid that if substitute

datawithdescr <- read.csv(file = full.name, sep = sep, 
            dec = dec, quote = "", header = FALSE, strip.white = TRUE, 
            na.strings = c("NA", "ND", ".", 
                "", "Missing"), stringsAsFactors = FALSE)

to

datawithdescr <- read.csv(file = full.name, sep = sep, 
            dec = dec, quote = "'\"", header = FALSE, strip.white = TRUE, 
            na.strings = c("NA", "ND", ".", 
                "", "Missing"), stringsAsFactors = FALSE)

not sure about side effects yet

[Helmut01]

Hi Michael,

Why would one do that? The CSV files are already provided in /library/replicateBE/extdata/*.csv… Given , with the commentary header but get.data.R handles it.

• The rds files are named S3-objects (with the factorized data.frame within). It’s not our job to prevent all possible errors a user might make. Actually the code for export should by write.csv(rds01, 'rds01.csv', quote=FALSE, row.names=FALSE).
• However, I will assess your suggestion.

[Helmut01]

Hi Michael,

there is a side effect.
With write.csv(rds01, 'rds01.csv') row.names are exported. After line 120 str(data) gives for export/import of rds01:
'data.frame': 298 obs. of 7 variables:
$ NA : int 1 2 3 4 5 6 7 8 9 10 ...
$ subject : chr "1" "1" "1" "1" ...
$ period : chr "1" "2" "3" "4" ...
$ sequence : chr "RTRT" "RTRT" "RTRT" "RTRT" ...
$ treatment: chr "R" "T" "R" "T" ...
$ PK : chr "2285.96" "1955.82" "1345.94" "2856.24" ...
$ logPK : chr "7.734541" "7.578565" "7.204848" "7.957261" ...
The first column is named NA (which is not a character like the others). Hence, we need:
if (typeof(data[[1]]) == "integer") data <- data[, -1]