help with ksh/awk/sed script, random # of fields

Shell Programming and Scripting
1

Hello all, I'm working on an attendance callout script for a school district. I need to change our current layout for the vendor. Currently the data is in the form of:

studentid,period,building,

Heres a sample of some made up records:

500,1,30,
500,2,30,
500,3,30,
500,6,30,
7899,2,31,
9021,1,33,
9021,6,33,
907711,5,40,
907711,6,40,

I need to reformat this, omitting the header row, to look like the following:

500,1,2,3,6,30,
7899,2,31,
9021,1,6,33,
907711,5,6,40,

I've done stuff like this with awk in the past when the number of fields were a constant. Just not sure how to loop through this when the number of fields vary. The kids can have an unexcused in any number of periods throughout the day. Some of our buildings have over 9 periods configured for attendance purposes as well.

If anyone has some code to get me going in the right direction that would be great.

I will read the file first then do something like this to compare the current record with the next record to verify its the same kid. (beyond that I'm not sure how to proceed)

awk -F, '
$1 == lastid {
         do something
}
$1 != lastid {
        lastid=$1
        do something
}'

Thanks,

2 
awk -F, -v OFS=, '{ for (I=2; I<NF; I++)
                    {print $1, $I}
                  }' <<END |\
sort -kn1,1 -kn2,2 -t, -u |\
awk -F, 'NR==1{Save0=$1}
         Save0 == $1{Line=Line "," $2; next}
         Save0 != $1 {print Save0  Line;
                      Line="," $2;
                      Save0=$1;
                     }
         END{print Save0  Line;}'
500,1,30,
500,2,30,
500,3,30,
500,6,30,
7899,2,31,
9021,1,33,
9021,6,33,
907711,5,40,
907711,6,40,

Produced

500,1,2,3,6,30
7899,2,31
9021,1,6,33
907711,5,6,40
3

Thanks a lot awk, I'll get to give your code a go on Monday first thing.

4

Assuming your data is sorted:

awk -F, '
NR==1{s=$1 FS $2; b=$3; next}
b!=$3{print s FS b; s=$1 FS $2; b=$3; next}
{s=s FS $2}
END{print s FS b}' file

Regards

5 
nawk -F"," '{
if (_[$1","$3]==""){
  _[$1","$3]=$2
  next
}
_[$1","$3]=sprintf("%s,%s",_[$1","$3],$2)
}
END{
for (i in _){
 split(i,arr,",")
 print arr[1]","_,","arr[2]
} 
}' filename
6

Thanks awk, works great for me. I was wondering what the <<END is in the first awk command? Is it like a here doc? I'm trying to understand this better to figure out what order the code is being executed in.

thanks again

7

Ideally, I should have had the word END on a line by itself after the data. It shows the korn shell where to stop the input of data to the awk program.

Still the script terminated at that point, the shell script figures out the END is implied.

Sorry, I will remember to put it in future posts.

8

heck, np at all. i'm just a newb so the only samples i've seen are pretty much in two books i have. thanks for getting back to me so fast.