How to process multiline records with awk

Blank-line-separated records can lose their context when a normal line-by-line search prints only the matching line. Use awk paragraph mode when a status, error, or field can appear inside a record that spans several lines and the output must keep the full record together.

Setting RS="" makes awk read records separated by one or more blank lines instead of one input line at a time. Setting ORS="\n\n" adds the blank line back between printed records so each matched block remains readable in terminal output or a handoff file.

The sample uses simple health-check blocks separated by empty lines. It is not a parser for JSON, YAML, quoted CSV, or application logs whose record boundary depends on nesting or escaping; use a format-aware parser for those files.

1. Create a sample file where each record spans several lines and records are separated by a blank line.

   $ cat > checks.txt <<'EOF'
   host: web-01
   status: ok
   service: nginx
   message: accepting requests
   
   host: db-01
   status: failed
   service: postgresql
   message: disk space below threshold
   
   host: cache-01
   status: ok
   service: redis
   message: responding to pings
   EOF

2. Print the whole record that contains the failed status.

   $ awk 'BEGIN { RS=""; ORS="\n\n" } /status: failed/' checks.txt
   host: db-01
   status: failed
   service: postgresql
   message: disk space below threshold

   RS="" switches awk to paragraph-style records. The pattern still searches within the current record, but print writes the entire matched record instead of only the matching line.

3. Confirm that the output includes the host, service, and message lines from the same failed record. The web-01 and cache-01 records should not appear because they do not contain status: failed.
4. Change the match expression for the field or text that identifies the real record. For example, use /service: postgresql/ when the service line is the decisive field, or combine record tests inside the awk program when two lines must both appear in the same block.

   Keep a real blank line between records when using RS="". If blank lines can appear inside a record, choose a record separator that matches the actual file format or use a parser built for that format.

5. Remove the sample file after testing.

   $ rm checks.txt