Awk and Pattern Match

[zen2003]

This is my data file:
UNA:+.? 'UNB+UNOA:3+CHASE:ZZ+ADSD:ZZ+030718:0103+00000000000630'UNH+0000
0000000326+CRL:2:2:UN'UCI+47000618072003+EV1:ZZ+CH:ZZ+7'UNT+3+000
00000000326'UNZ+1+00000000000630'
UNA:+.? 'UNB+UNOA:3+ss+WER:ZZ+030718:0110+00000000000631++BTS'
UNH+00000000012689+BTS:96A:UN'BGM+XZ8+1'DTM+137:20030718:102'LIN+1'RFF+XC3:
F2003-07-18-06.00.47.168473'RFF+AEK:03071806000002'RFF+CR:EXM 1000735AT1ZD'SE
Q++1'GIS+1'UNT+10+00000000012689'UNH+00000000012690+BTS:96A:UN'BGM+XZ8+2'DT
M+137:20030718:102'LIN+1'RFF+XC3:F2003-07-18-06.00.47.168473'RFF+AEK:03071806
000002'RFF+CR:EXM 1000736P139G'SEQ++1'GIS+1'UNT+10+00000000012690'UNH+0000000001
2691+BTS:96A:UN'BGM+XZ8+3'DTM+137:20030718:102'LIN+1'RFF+XC3:F2003-07-18
-06.00.47.168473'RFF+AEK:03071806000001'RFF+CR:EXM 1000939I361H'SEQ++1'GIS+1'UNT
+10+00000000012691'UNH+00000000012692+BTS:96A:UN'BGM+XZ8+4'DTM+137:20030718
:102'LIN+1'RFF+XC3:F2003-07-18-06.00.47.168473'RFF+AEK:03071806000003'RFF+CR:
FUS 1000312ASBCA'SEQ++1'GIS+1'UNT+10+00000000012692'UNH+00000000012693+BTS:
96A:UN'BGM+XZ8+5'DTM+137:20030718:102'LIN+1'RFF+XC3:F2003-07-18-06.00.47.1684
73'RFF+AEK:03071806000003'RFF+CR:FUS 1000313BTJRA'SEQ++1'GIS+1'UNT+10+0000000001
2693'UNH+00000000012694+BTS:96A:UN'BGM+XZ8+6'DTM+137:20030718:102'LIN+1'RFF
+XC3:F2003-07-18-06.00.47.168473'RFF+AEK:03071806000004'RFF+CR:FUS 1006351A1Y
AG'SEQ++1'GIS+1'UNT+10+00000000012694'UNH+00000000012695+BTS:96A:UN'BGM+XZ8
+7'DTM+137:20030718:102'LIN+1'RFF+XC3:F2003-07-18-06.00.47.168473'RFF+AEK:030
71806000004'RFF+CR:FUS 1006352FM3CB'SEQ++1'GIS+1'UNT+10+00000000012695'UNH+00000
000012696+BTS:96A:UN'BGM+XZ8+8'DTM+137:20030718:102'LIN+1'RFF+XC3:F2003-
07-18-06.00.47.168473'RFF+AEK:03071806000004'RFF+CR:FUS 1006353MC65A'SEQ++1'GIS+
1'UNT+10+00000000012696'UNH+00000000012697+BTS:96A:UN'BGM+XZ8+9'DTM+137:200
30718:102'LIN+1'RFF+XC3:F2003-07-18-06.00.47.168473'RFF+AEK:03071806000004'RF
F+CR:FUS 1006354Q29AA'SEQ++1'GIS+1'UNT+10+00000000012697'UNH+00000000012698+BANS
TA:96A:UN'BGM+XZ8+10'DTM+137:20030718:102'LIN+1'RFF+XC3:F2003-07-18-06.00.4
7.168473'RFF+AEK:03071806000004'RFF+CR:FUS 1006355R175B'SEQ++1'GIS+1'UNT+10+0000
0000012698'UNH+00000000012699+BTS:96A:UN'BGM+XZ8+11'DTM+137:20030718:102'LI
N+1'RFF+XC3:F2003-07-18-06.00.47.168473'RFF+AEK:03071806000004'RFF+CR:FUS 100
6356V353U'SEQ++1'GIS+1'UNT+10+00000000012699'UNZ+11+00000000000631'

what I need to do is get out all text that starts with - UNH
and ends with - UNT
so the script should return
UNH+0000
0000000326+CRL:2:2:UN'UCI+47000618072003+EV1:ZZ+CH:ZZ+7'UNT

UNH+00000000012689+BTS:96A:UN'BGM+XZ8+1'DTM+137:20030718:102'LIN+1'RFF+XC3:
F2003-07-18-06.00.47.168473'RFF+AEK:03071806000002'RFF+CR:EXM 1000735AT1ZD'SE
Q++1'GIS+1'UNT

and so on

Thanks for your help

 

[vgersh99]

if you're on Sun/Solaris, use 'nawk' istead of plain-old-broken 'awk'.

vlad
+----------------------------+
| #include<disclaimer.h> |
+----------------------------+

 

[zen2003]

I am using HP-UNIX 11

 

[PHV]

Try this:

BEGIN {
  RE="UNH.*"
  RS=FS="UNT"
}
{
  for(i=1;i<= NF;i++)
    if (match($i, RE))
      printf &quot;%s%s\n&quot;,substr($i,RSTART),FS
}

Hope This Help
PH.

 

[zen2003]

It does not come back with anything:

$ awk -f a5.awk contrl0.txt
+ awk -f a5.awk contrl0.txt
$

a5.awk is:
$ cat a5.awk
+ cat a5.awk
BEGIN {
RE=&quot;UNH.*&quot;
RS=FS=&quot;UNT&quot;
}
{
for(i=1;i<= NF;i++)
if (match($i, RE))
printf &quot;%s%s\n&quot;,substr($i,RSTART),FS>>&quot;p.txt&quot;
}

 

[vgersh99]

PHV,
that's a better idea to break down input by the 'UNT' records. It might work, but..... zen might also be hitting awk's limits that are RECORD-based.

Zen, if PHV's ssuggestion doesn't fix your problem, can you get your hands on either POSIX-compliant awk OR on gawk?

I'm not quite sure what HPUX is being shipped with.

vlad
+----------------------------+
| #include<disclaimer.h> |
+----------------------------+

 

[zen2003]

Nope, Can not get to either. I think I am hitting the awk's limits that are RECORD-based.If I reduce the amount of text then it works fine. Any ideas of how I can workaround this limit?

 

[PHV]

man fold

Hope This Help
PH.

 

[zen2003]

fold does not work as the record may be a variable lenght.
What I am trying to do is insert a newline before 'UNH'
This substitution works
sed 's/UNH/*!!@UNH/' contrl0.txt

What I want to do is
sed 's/UNH/geta newline char hereUNH/' contrl0.txt
I tried
sed 's/UNH/\\nUNH/' contrl0.txt and a few other combinations but it does not work

Can this be done?

Thanks

 

[PHV]

Try something like this:

sed 's/UNH/@UNH/' contrl0.txt | tr '@' '\012'

Hope This Help
PH.

 

[Ygor]

Try

sed 's/UNH/\^JUNH/' contrl0.txt

where ^J is Ctrl V Ctrl J