Hi,

I'm using the join command and it appears to discard certain fields. Here are the two files i'm comparing:

File1:
1 a
2 b
3 c
4 d
99 f
101 g
999 i
200 j

File 2:
1 e
2 f
3 g
4 h
99 h
101 i
999 j
200 k

> join 1 2
results in:

1 a e
2 b f
3 c g
4 d h
99 f h
999 i j

So 101 and 200 have been ignored....any ideas why?

Cheers

The files must be sorted first.

From the join manual page....
So....
Cheers
ZB

Thanks - you were right it was because it wasn't correctly sorted.

Does anyone know how to add a tab to the output of the join command i.e. the fields that are added should be seperated from the rest with a tab?


e.g.

1 g </tab> h
3 g </tab> i

Cheers

Well here's my join issue

I have a file in the following format:
id/name/salary/position

and a file in the following format
number,amount,id

if I use sed to replace the different separators with a ":" and try to join them - the join fails to match on id.

However if I type the files in with the : separator already in place - the join works fine.

Sadly I am the instructor - and tested the script with my pre-made files (with the ":") but assigned the work with different separators to make the class use sed to convert them.

Both files are sorted (ascii) and other than the separators are identical. Anyone have a clue as to why this would happen?

looking at 'man join':

I did that:

join -t: -1 3 -2 1 file1 file2

and if I hand type the files in - no problem. If however I use the original files with the differing separators, use sed to convert them to a ":" then the join fails. Same command, same file (except the separators in one have been manually entered, and the second time they were converted) different results

sigh such a pita

could you post a sample script AND sample input files?