Custom sorting for job columns

Deadline
1

Is there a way to control the sorting function of extra attribute columns?

2

Hey sir! How do you mean? Like customizing the sorting behaviour of the data itself?

Right now, I’d say “no” since it’s now all in C++ thanks to Jon’s great speed improvements of 2016.

Can you flesh that out a bit more and I can take something back to Jon?

3

We have a column populated by the estimated amount of data the job would be pulling off the network. This information is presented in a human readable form, like 55GB, or 1.4TB etc. Currently its sorting alphabetically like this:

1.0gb
10.98gb
102gb
1074gb
12gb
19gb
2gb
2.5gb
20gb
203gb
21gb
3gb
30gb

With a custom sorting function we could handle this appropriately.

Not a big deal, and i would certainly not give up the speed improvements for this :slight_smile:

4

As a hack workaround, you could create a sibling column with data structured specifically for sorting. The values in this column would need to follow a fixed decimal format that is large enough to cover all scenarios. The user workflow would be to click on the sibling column for sorting, but to read from the human-readable column.

 HUMAN      SORT (GB)
1.0gb       00001.0
10.98gb     00010.9
1.4tb       01400.0
102gb       00102.0
1074gb      01074.0

This could be accomplished in a single column if you don’t mind have the sorting data on the left:

     COMBINED
00001.0  [ 1.0gb ]
00010.9  [ 10.98gb ]
01400.0  [ 1.4tb ]
00102.0  [ 102gb ]
01074.0  [ 1074gb ]

Less than ideal, but perhaps better than nothing.

5

Good idea, thanks James!