breaking a file based on the Key column

[bengalliboy]

Hi
What would be the best way to split a file into pieces based on the 1st column (say 8 char, fixed length), some thing like this: The text file file1 looks like this:

ABCD001 asfafggfdggfgfg
ABCD001 hjhjhjh
ABCD001 hghgh
ABCD002 hjhjhj
ABCD002 jhhj

So I want: file2
ABCD001 asfafggfdggfgfg
ABCD001 hjhjhjh
ABCD001 hghgh

and file2:
ABCD002 hjhjhj
ABCD002 jhhj
TIA

 

[vgersh99]

using nawk. Save the below in myAwk.awk and run it as:

nawk -f myAwk.awk myTextFile.txt

#----------------------------- myAwk.awk--------------------

{
pos=match($1, "[0-9][0-9]*"
outFile= (!pos) ? "file0" : "file" substr($1, RSTART, RLENGTH);

print >> outFile;
}

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

 

[CaKiwi]

Use the following awk script

{
  if (substr($0,1,8) != hldstr) {
    if (NR>1) close(fn)
    fnix++
    fn = "file" fnix
    hldstr = substr($0,1,8)
  }
  print > fn
}

Put in a file, split.awk say, and enter

awk -f split.awk inputfile

CaKiwi

 

[vgersh99]

ooops, sorry - that's better

#----------------------------- myAwk.awk--------------------

{
pos=match($1, &quot;[1-9][0-9]*&quot;
outFile= (!pos) ? &quot;file0&quot; : &quot;file&quot; substr($1, RSTART, RLENGTH);

print >> outFile;
}

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

 

[vgersh99]

CaKiwi,
not sure if your implementation works:

ABCD001 asfafggfdggfgfg
ABCD001 hjhjhjh
ABCD001 hghgh
ABCD002 hjhjhj
ABCD003 vlad
ABCD002 jhhj

The file number is encoded in the value of the first column.

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

 

[CaKiwi]

Vlad,

I took it to mean that when the first 8 characters of a record changed the data should be written to a new sequentially numbered file, but your interpretation seems more likely. CaKiwi

 

[vgersh99]

whatever makes the customer happy vlad
+---------------------------+
|#include<disclaimer.h> |
+---------------------------+

 

[bengalliboy]

Thanks a lot... this is exactly what I am trying to do. Cawiki, I want to ask you if I can change the output file names from the file1, file2 etc to: first column itself... that is: file1 shold be named as ABCD001, fle 2 as ABCD002 ?
Thanks again...

 

[CaKiwi]

Yes, like this

.​

{
  if (substr($0,1,8) != hldstr) {
    if (NR>1) close(fn)
    fn = $1
    hldstr = substr($0,1,8)
  }
  print > fn
}

CaKiwi

 

[vgersh99]

CaKiwi,

you're a better mind-reader

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