UNIX for Dummies Questions & Answers

Dear all,

I 'd like to ask a question. I have two datasets: a.txt (only has one filed, call 'SNP'), b.txt( has thousands of fields, 1st field call 'SNP').

a.txt:
rs9527
rs318567
rs12376
...
b.txt:
rs167893 1 2 0 2 1 2 ...
rs318567 2 0 2 1 2 0 ...
rs12376 0 2 0 2 1 2 ...

I just want to generate dataset c.txt which is a subset of b.txt with 'SNP' names in a.txt
The code I can think of is: join -1 1 -2 1 -o 2.1 2.2 2.3 ... a.txt b.txt >c.txt

my question is: I want to keep all the fields in b.txt, while I have thousands of fields in b.txt. how to write the code ?

thanks a lot!

Lin

Why are you even bothering with any options to your join command? By default, it joins on the first field, so you don't need either -1 or -2. By default, the output is the join field followed by the rest of the fields from file1 (of which there are none in your sample data) and the rest of the fields of file2 (which is what you desire).

Do you need anything more complicated than join file1 file2?

By the way, your sample data isn't sorted. join requires sorted input files. If you are not using join's -t option, you must sort -b each file. If you are using -t, drop the -b. In either case, using the first field as the sort key.

Regards,
Alister

---------- Post updated at 05:08 PM ---------- Previous update was at 05:02 PM ----------

I just saw "subset". If it is too tedious to write the join -o command option manually, or if it exceeds the limits of your system even if generated dynamically, you can filter through cut.

You may be able to generate the option argument with a script, but you haven't provided any information about what's needed.

Regards,
Alister

thank you. as you said, the default works. I didn't know that before. thanks!
Lin