Different behavior of null_count when using regex field selector

[cmcgowanprovidertrust]

Hi there, I am encountering some behavior that I do not understand when trying to use stats1 -a null_count. The accumulator does not seem to work if I specify my columns using --fr instead of -f.

I am specifically working with pipe-delimited files and field names with spaces in them. I am not sure that matters at all.

Here is a simple example:

% cat testnull.csv 
Id|Type|Other Type|Number
a||foo|1
b||bar|2
c|||1

% mlr --oxtab --icsv --ifs "|" cat then stats1 -a null_count -f "Id","Type","Other Type","Number" testnull.csv
Id_null_count         0
Type_null_count       3
Other Type_null_count 1
Number_null_count     0

% mlr --oxtab --icsv --ifs "|" cat then stats1 -a null_count --fr ".*" testnull.csv                           
Id_null_count         0
Type_null_count       0
Other Type_null_count 0
Number_null_count     0

I would expect the call using --fr to produce the same results as the first, but it does not.

Apologies if this is just my misunderstanding of the tool and this is correct behavior, I am new to Miller and could be missing something.

[cmcgowanprovidertrust]

Version is 6.12.0, running in Zsh on Mac, installed with Homebrew.

[aborruso]

The same for me. It seems to me a bug cc @johnkerl