Feature request: missing data handling in plotting functions

[taprs]

Hi Wei, thank you for your great work!

It would be cool to make plotting functions work when there are rows with missing data. With csvtk v0.24.0 an error pops up when trying to plot a histogram with missing data:

$ echo $'1,1\n1,2\n,2'
1,1
1,2
,2
$ echo $'1,1\n1,2\n,2' | csvtk plot hist -f 1
[ERRO] fail to parse data:  at column: 1. please choose the right column by flag -f (--data-field)

[shenwei356]

Added a new flag --skip-na to csvtk plot:

  --na-values strings    NA values, case ignored (default [,NA,N/A])
  --skip-na              skip NA values in --na-values

Rember to use -H for data with no header line:

echo $'1,1\n1,2\n,2' | csvtk plot hist -H -f 1 --skip-na | display

Binaries:

• csvtk_darwin_amd64.tar.gz
• csvtk_linux_amd64.tar.gz
• csvtk_windows_amd64.exe.tar.gz

[taprs]

Hmm, but why not making it a default behavior, at least for empty cells?

[shenwei356]

Hmm, that seems reasonable.