matrix

#!/bin/bash

# this script will take a space-separated file of the format:
# identifier value
# identifier value

# it will then store each 'value' in the array according to its identifier.
# so the lines: 'foo bar' 'foo baz' 'other thing' will be stored as
# values[foo,1]=bar; values[foo,2]=baz; values[other,1]=thing
# the second array 'counts' will keep track of how many values exist for each
# identifier. so counts[foo]=2; counts[other]=1.

# yes, it's 1-indexed. this is preference, but i find it easier to keep track of

###############################################################################

# set vars
file=/path/to/whatever

# requires bash >= 4.0. this declares the array as associative* (see notes)
declare -A values counts

# read the file line by line, splitting into fields based on space**
while read -r iden val; do
  # increment 'counts' for the identifier
  ((counts[$iden]++))
  # store the value in the array
  values[$iden,${counts[$iden]}]=$val
done


# iterate over each identifier, dump the values. format will be:
# iden1: val1 val2
# iden2: val1 val2 val3

# loop over each identifier
for iden in "${!counts[@]}"; do
  # print identifier
  printf '%s:' "$iden"

  # loop over 1 through counts[iden]
  for ((i=1; i<=${counts[$iden]}; i++)); do
    # print each value
    printf ' %s' "${values[$iden,$i]}"
  done

  # print newline
  echo
done

###############################################################################
# notes:
# *     http://mywiki.wooledge.org/BashFAQ/006
# **    http://mywiki.wooledge.org/BashFAQ/001