Ytukyg

I am trying to perform an array definition in bash using values of another array as the name of the new array and assign the values inside the array from another array which is dynamically populated.

Here is an example of the code so far:

adduser() {

  declare -i segment=$1

  segment_length=${#segment[@]}

  for (( a = 0; a < "${segment_length}"; a++ )); do

    data=($(cat $filename | grep -w ${segment[a]} | awk -F ";" '{print $1}' | tr 'n' ' ' | sed 's/^/ /g'))

    ${segment[a]}=($(echo "${data[*]}"))

  done

}



cat $filename | tail -n+2 | awk -F ";" '{print $2}' | awk '{ for (i=1; i<=NF; i++) print $i }' | sed 's/r//g' | sort -u > segments.txt

IFS=$'rn' GLOBIGNORE='*' command eval  'segments=($(cat segments.txt))'

for (( i = 0; i < ${#segments[@]}; i++ )); do

adduser ${segments[i]}

done

The objective is to dynamically populate arrays (data in one column of csv) with values from another column and then massively work on them.

The csv has the following format:

Header1;Header2

Value1;1 2 3

Value2;2 4 5

Take for example the value 2 from column Header2.
The objective is to dynamically create the array 2 with values Value1 and Value2:

2=( Value1 Value2 )

Testing the two answers provided:

I will continue here as comments are too short: Here is the result from the answer with a random file (same format as the example):

ivo@spain-nuc-03:~/Downloads/TestStackoverflow$ awk -F'[; ]' '/;[0-9] / { for (i=2; i<=NF; i++) printf "%s ", $1 > $i".txt" }' real.csv 

ivo@spain-nuc-03:~/Downloads/TestStackoverflow$ ll

total 68

drwxr-xr-x  2 ivo ivo  4096 Nov 27 17:16 ./

drwxr-xr-x 28 ivo ivo 32768 Nov 27 17:15 ../

-rw-rw-r--  1 ivo ivo    99 Nov 27 17:16 155.txt

-rw-rw-r--  1 ivo ivo   132 Nov 27 17:16 155?.txt

-rw-rw-r--  1 ivo ivo    99 Nov 27 17:16 2.txt

-rw-rw-r--  1 ivo ivo    66 Nov 27 17:16 2?.txt

-rw-rw-r--  1 ivo ivo   198 Nov 27 17:16 3.txt

-rw-rw-r--  1 ivo ivo    33 Nov 27 17:16 3?.txt

-rw-r--r--  1 ivo ivo  1369 Nov 27 17:14 real.csv

While with the answer above you get the following:

ivo@spain-nuc-03:~/Downloads/TestStackoverflow$ ./processing.sh real.csv



ivo@spain-nuc-03:~/Downloads/TestStackoverflow$ ll

total 112

drwxr-xr-x  2 ivo ivo  4096 Nov 27 17:25 ./

drwxr-xr-x 28 ivo ivo 32768 Nov 27 17:15 ../

-rw-rw-r--  1 ivo ivo   100 Nov 27 17:25 102.txt

-rw-rw-r--  1 ivo ivo   100 Nov 27 17:25 105.txt

-rw-rw-r--  1 ivo ivo    67 Nov 27 17:25 106.txt

-rw-rw-r--  1 ivo ivo    34 Nov 27 17:25 112.txt

-rw-rw-r--  1 ivo ivo   991 Nov 27 17:25 155.txt

-rw-rw-r--  1 ivo ivo   694 Nov 27 17:25 2.txt

-rw-rw-r--  1 ivo ivo   859 Nov 27 17:25 3.txt

-rw-rw-r--  1 ivo ivo    67 Nov 27 17:25 51.txt

-rw-rw-r--  1 ivo ivo    67 Nov 27 17:25 58.txt

-rw-rw-r--  1 ivo ivo    34 Nov 27 17:25 59.txt

-rw-rw-r--  1 ivo ivo    34 Nov 27 17:25 65.txt

-rw-rw-r--  1 ivo ivo    34 Nov 27 17:25 67.txt

-rw-rw-r--  1 ivo ivo    34 Nov 27 17:25 72.txt

-rw-rw-r--  1 ivo ivo    34 Nov 27 17:25 78.txt

-rw-rw-r--  1 ivo ivo    34 Nov 27 17:25 81.txt

-rw-rw-r--  1 ivo ivo    34 Nov 27 17:25 82.txt

-rwxrwxr-x  1 ivo ivo  1180 Nov 27 17:25 processing.sh*

-rw-r--r--  1 ivo ivo  1369 Nov 27 17:14 real.csv

ivo@spain-nuc-03:~/Downloads/TestStackoverflow$

Thanks to everyone for their participation!

All-bash using array -

declare -a set=()                          # set is an array

while IFS=';' read -r key lst              # read each line into these 2 splitting on semicolons

do [[ $key =~ Header* ]] && continue       # ignore the header

   read -a val <<< "$lst"                  # split the list of values to the right of the semicolon into an array

   for e in "${val[@]}"                    # for each of those

   do  case "${set[e]:-}" in               

       *$key*) : already here, no-op   ;;  # ignore if already present

           '') set[e]="$key"           ;;  # set initial if empty

            *) set[e]="${set[e]} $key" ;;  # add delimited if new

       esac

   done

done < csv                                 # reads directly from the CSV file

At this point the sets should be loaded as space-delimited strings into each element of set, indexed by the values on the line in the second column of the csv. To print them out for verification -

for n in "${!set[@]}"

do echo "$n: ${set[n]}"

done

Executed on the test csv content provided:

1: Value1

2: Value1 Value2

3: Value1

4: Value2

5: Value2

so set[4] is Value2, and set[2] is Value1 Value2. You can pull them from there to do whatever is needed.

No need for cat/tail/awk/grep/tr/sed/sort chains.

Does it need something more?

Finally I went for a simpler method according to the comments I got to simplify:

cat $1 | tail -n+2 | awk -F ";" '{print $2}' | awk '{ for (i=1; i<=NF; i++) print $i }' | sed 's/r//g' | sort -u > segments.txt

IFS=$'rn' GLOBIGNORE='*' command eval  'segments=($(cat segments.txt))'

for (( i = 0; i < ${#segments[@]}; i++ )); do

cat $1 | grep -w ${segments[i]} | awk -F ";" '{print $1}' | tr 'n' ' ' | sed 's/^/ /g' > ${segments[i]}.txt

done

rm segments.txt

And then just process the remaining txt files in the folder.

I would really like to see this done in the way I was going initially though as it is more suitable for more data...

If I'm reading it right, and assuming you are ok with having no proper newline on the files since you were explicitly squashing them out, this should do all the above in one awk call.

awk -F'[; ]' '/;[0-9] / { for (i=2; i<=NF; i++) printf "%s ", $1 > $i".txt" }' yourInputFile