search

posit-dev / positron

Positron, a next-generation data science IDE
https://positron.posit.co
Other
2.72k stars 82 forks source link

Improve null / NA distinction for R in data explorer #2613

Open wesm opened 7 months ago

wesm commented 7 months ago

Currently in the summary stats dropdown, the null count is denoted with "NA". For some users, "NA" is used as null (and treated as null when being exported by R, e.g. into Parquet files via Arrow), but some users may wish to distinguish the number of NA values vs. the number of null values.

I think we should avoid using "NA" to indicate null values/counts in Positron, and if we want to introduce an "NA count" as an optional feature that language runtimes can support (R would support it, Python would not), then we can add this as an optional statistic to be computed and shown conditionally if the language runtime has support for it.

petetronic commented 6 months ago

Could we italicize and mark grey? Paul can provide the color value.

jmcphers commented 6 months ago

I opened https://github.com/posit-dev/positron/issues/2860 to track the treatment of these values in the Waffle.

jthomasmock commented 6 months ago

I'm fine with referring to explicitly missing data (NULL, NA, None, etc) as Missing in the summary stats.