Stop wrangling multiline patterns with regex nightmares
Text Processing (Grep/Sed/Awk)
Stop wrangling multiline patterns with regex nightmares
🧩 The Challenge
Dealing with application logs that spill over multiple lines is a special kind of hell when you’re just trying to grep for a specific stack trace. Most tools break on newlines, so you end up staring at partial data and missing the context you actually need.
💡 The Fix
Use awk’s record separator feature to redefine what constitutes a single “line.” By changing the separator to a blank line or a specific timestamp pattern, you can treat multiline blocks as a single cohesive unit for processing.
awk -v RS='\n\n' '/ERROR/ && /database_connection_pool/' /var/log/app/service.log
⚙️ Why It Works
Setting the RS variable tells awk to treat double newlines as the record boundary instead of the standard single newline. This keeps your entire multi-line error block together, letting you scan for multiple keywords within the same logical chunk of data.
🚀 Pro-Tip: Use RS='[0-9]{4}-[0-9]{2}-[0-9]{2}’ if your logs start every entry with a date to force awk to treat every log entry as one record.
Linux Tips & Tricks | © ngelinux.com | 11/21/2026
