Populate column selector for datasets which are not ready yet

[guerler]

Resolve #3758 for column parameters. ping @blankenberg.

[nsoranzo]

You can replace the last 2 lines with

                this_column_list = list(map(str, range(1, MAX_DEFAULT_COLUMNS + 1)))

which takes around half the time.
Similarly for lines 1134-1135.

[guerler]

Thx!

[blankenberg]

Awesome! Thanks. Will check it out soon.

[blankenberg]

Couple of concerns here.

I have a tool that takes a tabular input file and creates a tabular output file where the initial metadata of the output file is populated with a copy from the input dataset (metadata_source; this is ok). When when the job finishes, the output columns are correctly set to their real values. But I currently can still not select the newly created column because this change doesn't consider dataset set, only metadata presence.

Hardcoding 999 columns seems like a bandaid at best. This does not allow selecting e.g. column 1005 from the still running dataset.

I think we need to allow users to manually enter values (like it used to) and then validate them before the tool run.

See the Add a column tool, followed by trying to select the new column with e.g. the join by specified field tool.

[guerler]

Alright, Ill be revisiting this issue.

[blankenberg]

I agree it is nice to have the dropdown, since in many cases the number of columns might not change, but I wonder if there is a smooth way to allow both?