UNIX for Dummies Questions & Answers

Hi,

I have 7 big flat files, each contains 24 million records which have been sorted by the first field delimitered by Ctrl B (002).

I want to join them together side by side, eg.

File A:

1^Ba^Bb
2^Bx^By
....

File B:
1^Bc^Bd
2^Bm^Bn
....

After merged, it should look like :

File C :
1^Ba^Bb^Bc^Bd
2^Bx^By^Bm^Bn
....

Can anyone tell me if I can do this under unix and how ?

I will appreciate to any advices

xlil

put literally a tab in sed as 'paste' put a tab as delimiter.

Code:

paste filea fileb | sed "s/   .//"

Thanks, bhargav.

Code:

www# cat filea fileb
aaa
bbb
ccc
ddd
eee
fff
www# paste filea fileb | sed "s/   .//"
aaa     ddd
bbb     eee
ccc     fff

xli, you might want to redirect to a file:

Code:

paste filea fileb | sed "s/   .//"  > filec

Thanks, Neo and bhargav,

furthermore, is it possible to join more than two files at one time ?

Xli

Code:

paste filea fileb filec | sed "s/   .//" > filed

There is a join command. The use of cntl-B makes this hard to use from the command line. But as a shell script....

#! /usr/bin/ksh
join -t^B a b > c
exit 0

will do what was requested in the first post. See "man join" for more info.

Great!. That's exactly what I want. Two more questions :

First, as all of my flat files are huge, each contains around 24 millions records. If I join them tegether in one command, is it possible running out of memory ? In other words, does the script you wrote to me consumes memory very much ?

Second, I presume that data of the flat files might conain tab, so can I use other character as delimiter eg. Ctrl B, how to define in sed command ?

I appreciate to your help.

xli