Merge non nil values from one column into other

[dan-corneanu]

Hello,
I have the following dataframe

#<RedAmber::DataFrame : 1166 x 4 Vectors, 0x000000000000693c>
     date       trading_period availability adjustment
     <string>          <uint8>     <double>   <double>
   0 2025-01-08             35          1.0      (nil)
   1 2025-01-08             36          1.0      (nil)
   2 2025-01-08             37          1.0      (nil)
   3 2025-01-08             38          1.0      (nil)
   4 2025-01-08             39          1.0      (nil)
   5 2025-01-08             40          1.0      (nil)
   6 2025-01-08             41          1.0      (nil)
   7 2025-01-08             42          1.0      (nil)
   8 2025-01-08             43          1.0      (nil)
   9 2025-01-08             44          1.0      (nil)
  10 2025-01-08             45          1.0      (nil)
  11 2025-01-08             46          1.0      (nil)
  12 2025-01-08             47          1.0      (nil)
  13 2025-01-08             48          1.0      (nil)
  14 2025-01-09              1          1.0        0.5
  15 2025-01-09              2          1.0        0.5
  16 2025-01-09              3          1.0        0.5
  17 2025-01-09              4          1.0        0.5
  18 2025-01-09              5          1.0        0.5
  19 2025-01-09              6          1.0        0.5
   : :                       :            :          :
1162 2025-02-01             45          1.0      (nil)
1163 2025-02-01             46          1.0      (nil)
1164 2025-02-01             47          1.0      (nil)
1165 2025-02-01             48          1.0      (nil)

and I would like to add a new column :adjusted_availability that
should take the value of :adjustment if not nil, otherwise the
value of :availability.

The result should look like

 #<RedAmber::DataFrame : 1156 x 5 Vectors, 0x00000000000069f0>
     date       trading_period availability adjustment adjusted_availability
     <string>          <uint8>     <double>   <double>              <double>
   0 2025-01-08             45          1.0      (nil)                   1.0
   1 2025-01-08             46          1.0      (nil)                   1.0
   2 2025-01-08             47          1.0      (nil)                   1.0
   3 2025-01-08             48          1.0      (nil)                   1.0
   4 2025-01-09              1          1.0        0.5                   0.5
   5 2025-01-09              2          1.0        0.5                   0.5
   6 2025-01-09              3          1.0        0.5                   0.5
   7 2025-01-09              4          1.0        0.5                   0.5
   8 2025-01-09              5          1.0        0.5                   0.5
   9 2025-01-09              6          1.0        0.5                   0.5
  10 2025-01-09              7          1.0        0.5                   0.5
  11 2025-01-09              8          1.0        0.5                   0.5
  12 2025-01-09              9          1.0        0.5                   0.5
  13 2025-01-09             10          1.0        0.5                   0.5
  14 2025-01-09             11          1.0        0.5                   0.5
  15 2025-01-09             12          1.0        0.5                   0.5
  16 2025-01-09             13          1.0        0.5                   0.5
  17 2025-01-09             14          1.0        0.5                   0.5
  18 2025-01-09             15          1.0        0.5                   0.5
  19 2025-01-09             16          1.0        0.5                   0.5
   : :                       :            :          :                     :
1152 2025-02-01             45          1.0      (nil)                   1.0
1153 2025-02-01             46          1.0      (nil)                   1.0
1154 2025-02-01             47          1.0      (nil)                   1.0
1155 2025-02-01             48          1.0      (nil)                   1.0

My solution was the following

df.assign(:adjusted_availability) { adjustment.map(&:nil?).not }
  .assign(:adjusted_availability) { adjusted_availability.if_else(adjustment, availability) }

Is there a more performant or more idiomatic way to achieve this result?

[kou]

This will work but we should add convenient API for :coalesce something like adjustment.coalesce(availability):

df.assign(:adjusted_availability) { Arrow::Function.find(:coalesce).execute([df.adjustment.data, df.availability.data]) }