AIX join command

[David Teh]

Hi guys,

Trying to get something to work here with the unix join command.

I have an IN.dat file (1st column is MSG ID and 2nd column is UNIX TIME):

6839944 1251732690

6839945 1251732709

6839946 1251732722

6839947 1251732722

6839948 1251732735

(etc)

I have an OUT.dat file:

6958611 1246807903

6958612 1246808006

6958613 1246808090

6958614 1246808115

6958615 1246808259

6958616 1246808337

6958617 1246808434

6958618 1246808548

The records in the 2 files may not necessary match.

I basically want 4 columns in a 3rd file as shown:

“IN MSG ID”  “IN TIME” “OUT MSG ID” “OUT TIME”     [if matching]

(none) (none) “OUT MSG ID” “OUT TIME”     [if not matching]

“IN MSG ID”  “IN TIME” (none) (none)      [if not matching]

The following command seems to have worked for others but not me:

join -t “,” -a 1 -a 2 -e ‘(none)’ -o 0,1.2,2.2 IN.dat OUT.dat

Output I got:

6839944 1251732690,(none),(none)

6839945 1251732709,(none),(none)

6839946 1251732722,(none),(none)

6839947 1251732722,(none),(none)

6839948 1251732735,(none),(none)

6839949 1251732783,(none),(none)

6839950 1251732783,(none),(none)

6839951 1251732802,(none),(none)

6839952 1251732833,(none),(none)

6839953 1251732881,(none),(none)

6839954 1251732894,(none),(none)

6839955 1251732925,(none),(none)

6839956 1251732979,(none),(none)

6839957 1251733009,(none),(none)

“6839944 1251732690” (no comman??) seem to suggest that the command treats it as a single field instead of 2 fields?

Any idea what’s wrong?

TIA!

Topic

[Abe Rastkar]

It seems like since you used a comma as the field separator, it picked up both fields as one. the -t represents a separator in the input and the output files. The following is from the man page

-t Character  Uses the character specified by the Character parameter as the field separator character in the input and the output. Every appearance of the character in a line is significant. The default separator is a space. With default field separation, the collating sequence is that of the sort -b command. If you specify -t, the sequence is that of a plain sort. To specify a tab character, enclose it in single quotation marks.

[David Teh]

thought -t was for the resulting output…let me test it out first…

thanks!

Abe Rastkar wrote:

It seems like since you used a comma as the field separator, it picked up both fields as one. the -t represents a separator in the input and the output files. The following is from the man page

-t Character

[David Teh]

reporting back…….

“problem” with join is that you need to do a sort on both files first on the key column…..

[Author]

Replies