topic Re: column selection/data subset in Discussions

column selection/data subset

abbasintx — Thu, 13 Jun 2024 18:45:40 GMT

Need a little help. In my data table i have like 500 columns and 2000 rows. I have another data table which contains only a small ## of column names (subset of columns of the original table). How do i subset my my original data table so it contains only the columns which are i have a list for in the second table.?

Re: column selection/data subset

jthi — Thu, 13 Jun 2024 19:05:53 GMT

I think this interactive flow should work:

I have two tables, this small table determines the columns I wish to keep

My other table has much more data but I need to just get those columns

you can use Join to basically overwrite the values is smaller table

With pure script solution it could look something like this. First get the column names from the smaller table into a list using << Get Column Names("String") and then use that list in << Subset()

Names Default To Here(1);

dtbig = Open("$SAMPLE_DATA/Big Class Families.jmp");

dtsmall = Open("$SAMPLE_DATA/Big Class.jmp");

cols_to_keep = dtsmall << Get Column Names("String");

dt_new = dtbig << Subset(All Rows, Columns(cols_to_keep));

Re: column selection/data subset

abbasintx — Thu, 13 Jun 2024 21:27:06 GMT

Thanks...i think stacking and un stacking does it for me...Appreciate your help..

Re: column selection/data subset

txnelson — Thu, 13 Jun 2024 21:50:47 GMT

Just a comment about stacking. By stacking and splitting(un stacking) the column metadata is lost, where using either of the techniques that Jarmo suggested does not. Formats, formulas, Column Properties such as Units, Value Labels, Spec Limits, Control Limits will be lost.

Re: column selection/data subset

abbasintx — Fri, 14 Jun 2024 00:19:03 GMT

Ah… I think join works the best few clicks…you all are amazing.