In-line replacement bash (replace line with new one using variables) - linux

Question

I'm going through and reading lines from a file. They have a ton of information that is unnecessary, and I want to reformat the lines for later use so that I can use the necessary information later.
Example line in file (file1)
Name: *name* Date: *date* Age: *age* Gender: *gender* Score: *score*
Say I want to just pull gender and age from the file and use that later
New line
*gender*, *age*
In bash:
while read line; do
<store variable for gender>
<store variable for age>
<overwrite each line in CSV - gender,age>
<use gender/age as inputs for later comparisons>
done < file1
EDIT: There is no stability in the entries. One value can be found using a echo $line | cut and the other value is found using a [ $line =~ "keyValue" ] then setting that value
I was thinking of storing the combination of the two variables as such:
newLine="$val1,$val2"
Then using a sed in-line replace to replace the $line with $newLine.
Is there a better way, though? It may come down to a sed formatting issue with variables.

Answer 1

This will produce your desired output from your posted sample input:
$ cat file
Name: *name* Date: *date* Age: *age* Gender: *gender* Score: *score*
$ awk -F'[: ]+' -v OFS=', ' '{for (i=1;i<NF;i+=2) a[$i]=$(i+1); print a["Gender"], a["Age"]}' file
*gender*, *age*
$ awk -F'[: ]+' -v OFS=', ' '{for (i=1;i<NF;i+=2) a[$i]=$(i+1); print a["Score"], a["Name"], a["Date"] }' file
*score*, *name*, *date*
and you can see above how easy it is to print whatever fields you like in whatever order you like.
If it's not what you want, post some more representative input.

Answer 2

Your example leaves room for interpretation, so I'm assuming that there may be whitespace in the field values, but that there are no colons in the field values and that each field key is followed by a colon. I also assume that the order is stable.
while IFS=: read _ _ _ age gender _; do
age="${age% Gender}" # Use parameter expansion to strip off the key for the *next* field.
gender="${gender% Score}"
printf '"%s","%s"\n' "$gender" "$age"
done < file1 > file1.csv
Update
Since your question now states that there is no stability, you have to iterate through the possible values to get your output:
while IFS=: read -a line; do
unset age key sex
for chunk in "${line[#]}"; do
val="${chunk% *}" # Everything but the key
case "$key" in
Age) age="$val";;
Gender) sex="$val";;
esac
# The key is for the *next* iteration.
key="${chunk##* }"
done
if [[ $age || $sex ]]; then
printf '"%s","%s"\n' "$sex" "$age"
fi
done < file1 > file1.csv
(Also I added quotes around the output values in the csv to be compliant with the actual csv format and in case sex or age happened to have commas in it. Maybe someone is 1,000,000 years old. ;)