I've been trying so solve this problem but none of the things I tried actually work. I have a table in which (from a .txt file) that I need to sort by numerical order of a specific raw of the table.
The data I have is composed like this:
Name Team Age
David a 20
James b 25
Anthony c 22
Carter b 21
Jessica a 24
Basically what I need to order (from smallest to biggest) the age of a certain team, example A. So the expected output would be:
a 20
a 24
awk '$2==team{print $2,$3}' team='a' file | sort -t' ' -k2n
-r separately if it doesn't work for you.sort --version?You could handle this very easily with SQL and the SUM and GROUP BY functions:
SELECT team, SUM(age) FROM table1 GROUP BY team;
+------+----------+
| team | SUM(age) |
+------+----------+
| a | 44 |
| b | 46 |
| c | 22 |
+------+----------+
Bash alone is clearly not the best tool for this job.
If you love Bash code like I do, you will probably still be interested in a Bash-only solution (well, almost, let's not reinvent the sorting):
#!/bin/bash
firstline=true
declare -A sumPerTeam # declare associative array
# collect sum for each team in the associative array:
while read name team age
do
[ "$firstline" = true ] && { firstline= ; continue ; } # skip header
sumPerTeam[${team}]=$(( "${sumPerTeam[${team}]}" + age ))
done < input.txt
# output each key/value pair of the array:
for team in "${!sumPerTeam[@]}"
do
echo "$team" "${sumPerTeam[${team}]}"
done | sort -k2
Associative arrays are not Bash's most readable feature.
An awk solution without a separate pipe and subshell to sort could be:
awk '$2 == "a" {arr[$3]} END {for (i in arr) print "a", i}' file
You simply store the age as indexes in an array and then output the indexes at the end which provides an inherent sort without a separate process, e.g.
Example Use/Output
$ awk '$2 == "a" {arr[$3]} END {for (i in arr) print "a", i}' file
a 20
a 24
Calling an external sort, you can reduce the number of characters you type, but then pickup the additional overhead of the pipe as well as the subshell required by sort, e.g.
awk '$2 == "a" {print $2, $3}' file | sort
(same output)
a 10 will come before a 2.