Can I combine contents of 2 files into a third file ?

mq15 · 02-06-2010, 04:22 AM

Hi there,
Suppose I have 2 files:

one.dat

Code:

common      one
1           20
2           30
30          40
...         ...
n            m

two.dat

Code:

common      two
1           10
2           15
30          20
...         ...
n            p

Can I use awk or something els like awk to get a third file (say one-cum-two.dat) with contents like this:

one-cum-two.dat

Code:

common      one     two
1           20      10
2           30      15
30          40      20
...         ...     ...
n            m       p

Kindly guide me. Best Regards

sohail0399 · 02-06-2010, 04:40 AM

may be this can help you

cat one.dat && tail -n "`wc -l two.dat | awk '{ print $1-1}'`" two.dat

mq15 · 02-06-2010, 07:34 AM

Thanks sohail0399
When I did as you said, this is what happened:

Code:

[mq15@localhost try]$ cat one.dat && tail -n "`wc -l two.dat | awk '{ print $1-1}'`" two.dat
common      one
1           20
2           30
30          40
...         ...
n            m
1           10
2           15
30          20
...         ...
n            p
[mq15@localhost try]$

I want the output to be the same as one.dat but with a third column (i.e two) from two.dat file.

pixellany · 02-06-2010, 08:39 AM

Uncle Google found the "paste" command:

Code:

[mherring@Ath play]$ more one
common      one
1           20
2           30
30          40

[mherring@Ath play]$ more two
common      two
1           10
2           15
30          20

[mherring@Ath play]$ paste one two|awk '{printf("%s\t%s\t%s\n", $1,$2,$4)}'
common  one     two
1       20      10
2       30      15
30      40      20

jschiwal · 02-06-2010, 09:21 AM

There is also the join command:

Code:

cat one                      
1       20                                               
2       30
3       4
30      40
35      45
40      50
n       m
> cat two
1       23
2       33
3
30      43
33
40      53
n       o
> join -j1 one two -e# -t'    '
1       20      23
2       30      33
3       4       #
30      40      43
40      50      53
n       m       o

The character after the -t option is a tab. The input files need to be sorted on the join fields.

mq15 · 02-06-2010, 10:25 AM

Thanks to everyone here..

pixellany · 02-06-2010, 12:07 PM

Quote:

Originally Posted by jschiwal

There is also the join command:

Code:

cat one                      
1       20                                               
2       30
3       4
30      40
35      45
40      50
n       m
> cat two
1       23
2       33
3
30      43
33
40      53
n       o
> join -j1 one two -e# -t'    '
1       20      23
2       30      33
3       4       #
30      40      43
40      50      53
n       m       o

The character after the -t option is a tab. The input files need to be sorted on the join fields.

In this method, the lines beginning with 33 and 35 are missing.

My method also fails when some lines are missing fields.

pixellany · 02-06-2010, 02:48 PM

I have put this back to "unsolved" status-----because 2 solutions (including mine) have some issues when fields are empty.

A fix that's in my head is to first preprocess the files and put a word (eg "none") in place of the empty fields. While my head tries to do this, I fully expect an AWK wizard to jump in with something more elegant....