0

I've built an SQL migration script that will replace text in a MySQL dump file. Replacements would be easy with sed, but we must account for special serializations (PHP Serialize). The code below works but the awk command is very expensive. With a quick google search, I see many developers with the same runtime performance issues. This article breaks down how simple changes can make a large difference. If anyone knows how I can speed this command up, or faster alternatives, that would be amazing :)

#!/usr/bin/env bash

set -e

SQL_FILE="$1"

maindomain="$2"

localdomain="$3"

sed 's/;s:/;\ns:/g' "$SQL_FILE" | \
  awk -F'"' '/s:.+'$maindomain'/ {sub("'$maindomain'", "'$localdomain'"); n=length($2)-1; sub(/:[[:digit:]]+:/, ":" n ":")} 1' | \
  sed -e ':a' -e 'N' -e '$!ba' -e 's/\n/ /g' | \
  sed "s/$maindomain/$localdomain/g" > "$SQL_FILE.txt"

rm "$SQL_FILE"

mv "$SQL_FILE.txt" "$SQL_FILE"
Richard Tyler Miles
  • 103
  • 5

1 Answers1

0

My issue faced above was solved by updating the libraries behind awk on (MacOS 12.6 (21G115)). After running brew install cgrep and brew install gawk the awk -F'"' '/s:.+'$maindomain'/ {sub(... command shown in the problem statement above saw dramatic improvements. From 3-5 minutes to ~ sys 0m0.039s.

I did not need to change the awk command to cgrep or gawk.

Richard Tyler Miles
  • 103
  • 5