[SOLVED] reformatting a file with columns using AWK

atjurhs · 04-19-2024, 11:36 AM

hi guys,

i wrote an AWK script that prints out data using this command

Code:

printf("\n%s,%s,%.5f,%.5f,%s",aray1,aray2,aray3,aray4,aray5)

it gives me output that looks like

Code:

fieldname1,val11,fieldname2,val21,fieldname3,val31,fieldname4,val41,fieldname5,val51
fieldname1,val12,fieldname2,val22,fieldname3,val32,fieldname4,val42,fieldname5,val52
fieldname1,val13,fieldname2,val23,fieldname3,val33,fieldname4,val43,fieldname5,val53
fieldname1,val14,fieldname2,val24,fieldname3,val34,fieldname4,val44,fieldname5,val54
fieldname1,val15,fieldname2,val25,fieldname3,val35,fieldname4,val45,fieldname5,val55
fieldname1,val16,fieldname2,val26,fieldname3,val36,fieldname4,val46,fieldname5,val56

i need my output to look like

Code:

fieldname1,fieldname2,fieldname3,fieldname4,fieldname5
val11,val21,val31,val41,val51
val12,val22,val32,val42,val52
val13,val23,val33,val43,val53
val14,val24,val34,val44,val54
etc.

i don't know how to "transpose" the data. maybe it would be better to use a different kind of print statement in my AWK script? or maybe easier to write a another AWK script to do it, idk?

guidance would be much appreciated, thank you

blunix2 · 04-19-2024, 11:49 AM

Can you paste us some example input please, and the complete awk script

atjurhs · 04-19-2024, 12:25 PM

i would love to, but they are on a separate system that i can't get data on or off. i have to manually type back and forth between here and there.

MadeInGermany · 04-21-2024, 04:55 AM

Each arayN is a name,val tuple?
Then the following is a way:

Code:

function tupleprintf(fmt, tup1, tup2, tup3, tup4, tup5) {
# Split each tupN at the comma to ArN[]
  split(tup1, Ar1, ",")
  split(tup2, Ar2, ",")
  split(tup3, Ar3, ",")
  split(tup4, Ar4, ",")
  split(tup5, Ar5, ",")
  if (!contd) {
# First elements (names) make the header; print once
    printf fmt, Ar1[1], Ar2[1], Ar3[1], Ar4[1], Ar5[1]
    contd=1
  }
# Print second elements (values)
  printf fmt, Ar1[2], Ar2[2], Ar3[2], Ar4[2], Ar5[2]
}

...

{
#printf("\n%s,%s,%s,%s,%s",aray1,aray2,aray3,aray4,aray5)
# Print all as %s strings
tupleprintf("\n%s,%s,%s,%s,%s",aray1,aray2,aray3,aray4,aray5)
}

grail · 04-21-2024, 06:39 AM

I am curious, of the several hundred search responses, either on google or even just here in LQ, which ones did you try before deciding that your data was too unique for any of the solutions?

atjurhs · 04-22-2024, 06:08 AM

grail, i did not know what to search for. if you could give me some clues, i would certainly do the searching.

syg00 · 04-22-2024, 06:24 AM

I can sympathise with the OP - you almost need to know the answer to the question to be able to frame the query.

Albeit, the comment in post #2 are cogent - we can't provide assistance in the dark. Nice response from MadeInGermany as I have come to expect.

sundialsvcs · 04-22-2024, 06:45 AM

When I have to do things like this, I turn to the perl programming language. Which grew out of "awk" when a guy named Larry Wall decided that he could improve it. It is definitely a "quirky" language, but extremely powerful – especially for any sort of "text-file processing."

One important thing that this language has, which "awk" does not, is a vast(!) library of packages (called "CPAN") which do "most of the things that you really need to do" right out of the box.

pan64 · 04-22-2024, 06:54 AM

Quote:

Originally Posted by syg00

I can sympathise with the OP - you almost need to know the answer to the question to be able to frame the query.

The most important part is to split the task into simple steps.
print header, print values, loop on input ... print etc.
At the end you will get something similar to post #4, which is a relatively compact implementation.

atjurhs · 04-22-2024, 10:42 AM

pan64, that's what i did, worked great, thanks

grail · 04-24-2024, 10:54 AM

I know we have a solution, but thought I would throw an awk one in for the hell of it:

Code:

#!/usr/bin/awk -f

BEGINFILE { RS="[,\n]"
            PROCINFO["sorted_in"] = "@ind_str_asc"
}

FNR % 2 == 1 && FNR<10{
  header = (header?header",":"") $0
}

FNR % 10 == 1{ ++cnt }

FNR % 2 == 0{ arr[cnt] = (arr[cnt]?arr[cnt]",":"") $0 }

END{
  print header
  for(idx in arr)
    print arr[idx]
}

Tis can of course be done on one line, but used this format for some clarity