I have a table below in KDB+ that I need to reorder columns for, but not by name, by Index. What I mean is that in the example table below, I don't know that the column names are "ltCt", "poPt", "kgAg", "xoOt", all I know is that the 6th column needs to be the 4th column, the 7th column needs to be the 5th column, etc., and I need to reorder them according to that.
Is it possible to reorder columns like this in KDB+? My understanding is that xcols seems to require column names.
x y z ltCt poPt kgAg xoOt
---------------------------------------------------------------------
today setting1 model1 41.66 367.69 -662.89 11347.91
ideal output format:
x y z kgAg xoOt ltCt poPt
---------------------------------------------------------------------
today setting1 model1 -662.89 11347.91 41.66 367.69
You can use the keyword cols to extract the list of columns:
q)cols tab
`x`y`z`ltCt`poPt`kgAg`xoOt
edit: You can then use the keyword rank followed by the order you would like (or supply just the indexes without rank) to reorder the column list:
q)cols[tab] rank 1 2 3 6 7 4 5
`x`y`z`kgAg`xoOt`ltCt`poPt
q)cols[tab] 0 1 2 5 6 3 4
`x`y`z`kgAg`xoOt`ltCt`poPt
Then finally put this into xcols along with your table:
q)(cols[tab] 0 1 2 5 6 3 4) xcols tab