Closed khughitt closed 4 years ago
Greetings!
Is there any reason why the R implementation of read_feather() ignores the state of the stringsAsFactors option?
read_feather()
stringsAsFactors
For consistency, it seems like it would be useful to have it behave in a similar manner to read.delim(), read_tsv(), etc...
read.delim()
read_tsv()
Ex:
library(feather) library(tidyverse) options(stringsAsFactors = FALSE) write_tsv(head(iris), 'test.tsv') write_feather(head(iris), 'test.feather') head(read.delim('test.tsv', sep='\t')$Species) # [1] "setosa" "setosa" "setosa" "setosa" "setosa" "setosa" head(read_tsv('test.tsv', col_types = cols())$Species) # [1] "setosa" "setosa" "setosa" "setosa" "setosa" "setosa" head(read_feather('test.feather')$Species) # [1] setosa setosa setosa setosa setosa setosa # Levels: setosa versicolor virginica
Incidentally, I checked read_parquet() and it too ignores stringAsFactors.
read_parquet()
stringAsFactors
Versions:
Can you open an issue on the Arrow JIRA issue tracker?
Sure thing - reported issue here: https://issues.apache.org/jira/browse/ARROW-7823
I created a similar issue for read_parquet() as well.
Thanks!
Greetings!
Is there any reason why the R implementation of
read_feather()
ignores the state of thestringsAsFactors
option?For consistency, it seems like it would be useful to have it behave in a similar manner to
read.delim()
,read_tsv()
, etc...Ex:
Incidentally, I checked
read_parquet()
and it too ignoresstringAsFactors
.Versions: