R: Speedups in operations with regular expressions

[NewsBot]

A new R blog post/announcement has been posted!

Regular expression operations in R, such as grep or gsub, sometimes have significant performance overheads due to encoding conversions.
Some R code tries to mitigate this by ignoring input encodings and pretending it is fine to work on individual bytes (via useBytes=TRUE). This removes such overheads, but produces correct results only in special cases, e.g. for simple regular expressions in UTF-8. With the current implementation of gsub and strsplit in R, this can also silently introduce invalid strings, which may cause invalid results or errors in further processing.

Get the full details here: Speedups in operations with regular expressions - The R Blog

[bot]

Corresponding tweet for this thread:

Devtalk

Share link for this tweet.