drl,

I noticed you used sh in your script and I used bash when I did my interactive session. So I changed my shell to sh but still join did not work. Here is the session history:

[chenz@beta binned]$ echo $0
bash
[chenz@beta binned]$ sh
sh-3.1$ echo $0
sh
sh-3.1$ cat data1
rs10051507 5 q21.3
rs10051514 5 p15.32
rs10051527 5 q21.2
rs1005152 7 q21.3
rs10051540 5 q21.3
rs10051548 5 q21.1
rs1005155 X q27.3
rs10051594 5 q34
sh-3.1$ cat data2
rs1003456 3
rs1005152 3
sh-3.1$ sort data1>t1
sh-3.1$ sort data2>t2
sh-3.1$ join t1 t2
sh-3.1$

Hi.

Yes, I think the LC_ALL and LANG settings may have something to do with it. It looks like you are running RedHat 5.

Note that your default LANG is neither C nor null, both of which worked for me.

You can try entering this interactively:
and then enter the steps again.

I have no additional thoughts. Perhaps someone will stop by with some advice ... cheers, drl

As drl pointed out file1 and file 2 are not sorted. Sort both of 'em before trying the join operation to get the expected result.

Thanks to Shamrock -- it's working now using the following syntax:

sort -k1,1 file1 > file1srt

sort -k1,1 file2 > file2srt

join -j 1 file1srt file2srt

Out of curiosity: why the following not working?

sort -k1 file1 > file1srt

sort -k1 file2 > file2srt

join -j 1 file1srt file2srt

In any case, sincerely thanks to both drl and shamrock.

"sort -k1" sorts file on the entire line. The start field is 1 but as no end field is given it defaults to the end of the line. "sort -k1,1" specifies the start and end field to be 1 and so its output is different from the "sort -k1" case.

Thanks a lot, shamrock.