Can you help me please? I have a task. I have from input some text with numbers. For example:
beta 1
score 9
something 2
beta 4
something 1
I need to calculate all numbers with the same text around. And my output will be:(in this way with ":")
beta:5
something:3
score:9
Also it can be problem with temp files, where i can save my scores. And I need to use mktemp after script finished to delete it.
Help me please, thanks.
I will be assuming that the input will always contain exactly two fields per line.
You may use the GNU datamash utility to sort the data, group it by the first field, and calculate the sum of the second field for each group:
datamash -s -W --output-delimiter=: groupby 1 sum 2 <file
Here, the -s sorts the input, -W makes the utility treat any run of consecutive whitespace characters as a field delimiter, and --output-delimiter=: sets the output delimiter to the : character. The rest tells datamash to group by the first field and to calculate the sum of the second field for each group.
Given the input in the question in the file called file, this would produce the following output:
beta:5
score:9
something:3
You can solve this in any number of other ways too. The easiest computational solution would be to use awk:
awk '{ sum[$1] += $2 } END { for (key in sum) printf "%s:%d\n", key, sum[key] }' file
Here, we use an associative array, sum, to hold the sum for each of the strings in the first field. The END block executes at the end of the input and outputs the calculated sums together with the strings.
Note that this solution also assumes that the first field is a single word containing no whitespace characters, as shown in the question.
Using a shell loop, reading the sorted lines from the original file, printing and resetting the sum of the second field whenever a new first field is encountered:
unset -v prev
sort file |
{
while read -r key value; do
if [ "$key" != "${prev-$key}" ]; then
# prev is set and different from $key
printf '%s:%d\n' "$prev" "$sum"
sum=0
fi
prev=$key
sum=$(( sum + value ))
done
if [ "${prev+set}" = set ]; then
printf '%s:%d\n' "$prev" "$sum"
fi
}
Related: Why is using a shell loop to process text considered bad practice?
awk program (or, for that matter, encapsulated by the datamash command). You may possibly have to update your question if you have further clarifications to it.
If you are dealing with large file, consider using sort and awk so that we don't allocate huge array for storing key and values in the RAM.
λ cat input.txt
beta 1
score 9
something 2
beta 4
something 1
sort input.txt |
awk -v OFS=: 'NR==1{ key=$1 }; NR>1&&$1!=key{ print key, sum; sum=0; key=$1 }; {sum+=$2} END{ print key, sum}'
beta:5
score:9
something:3
#!/bin/bash
declare -i SECOND
while read first second; do
if [ -z $FIRST ] || [ $first = $FIRST ]; then
SECOND+=second
else
echo $FIRST:$SECOND
SECOND=second
fi
FIRST=$first
done < <(sort file)
echo $FIRST:$SECOND
Usually I write a similar blank and in production I put all the variables in quotes.
for k in $(awk '{if(!seen[$1]++)print $1}' file.txt); do awk -v k="$k" 'BEGIN{sum=0}$0 ~ k {sum=sum+$2}END{print k,sum}' file.txt; done
output
beta 5
score 9
something 3
file.txt, comprising n*n lines of data. For comparison all the other solutions are O(n). Obviously for a homework exercise such as this with small values of n (lines in the file) there'll be little difference, but for a larger file such as a 1000 lines, your solution would read the entire file of 1000 lines up to 1001 times.
mktempcreates a file, it does not delete anything. Also, it seems that there is no need for temporary files to solve this exercise.