Unix ksh脚本验证：统计多日志中PID出现次数是否可行

Hey there! Let's walk through validating your Ksh script for counting PID occurrences across abc.log and xyz.log—and also cover a solid reference implementation so you can cross-check your work.

You need to:

• Count total occurrences of each PID across both log files (duplicates within a single file and cross-file repeats should all contribute to the PID's total count)
• Handle logs where PIDs follow a pattern like PID:6543 (assuming this is consistent across your log entries)

To confirm your script works as intended, verify these critical details:

1. It processes both log files: Ensure your script reads from both abc.log and xyz.log—either by passing them as arguments, or explicitly including them in your command chain. For example, if using grep or awk, you need to list both files as inputs.
2. PID extraction is accurate: If your logs use the PID:[number] format, your script should reliably pull just the numeric PID value. For instance, using sed 's/.*PID://' to strip text before PID:, or awk to directly extract the relevant segment.
3. It aggregates counts correctly: You need to sum occurrences across both files, not just count per file. Tools like awk are ideal here because they can track counts in memory as they process lines from both inputs.
4. It handles edge cases: Make sure lines without a PID are skipped instead of counting empty/non-numeric values that could skew results.

If you want to compare your script to a proven working version, here's a concise Ksh-compatible command chain that meets your requirements:

awk '/PID:/ { split($0, arr, "PID:"); pid = arr[2]; gsub(/[^0-9]/, "", pid); count[pid]++ } END { for (p in count) print p, count[p] }' abc.log xyz.log | sort -n

Let’s break this down:

• awk '/PID:/': Only processes lines that contain the PID: pattern (skips irrelevant lines)
• split($0, arr, "PID:"): Splits each matching line into an array using PID: as the delimiter
• pid = arr[2]; gsub(/[^0-9]/, "", pid): Grabs the text after PID: and removes any non-numeric characters (in case extra text follows the PID)
• count[pid]++: Increments the count for that PID in an associative array
• END { for (p in count) print p, count[p] }: After processing all lines, prints each PID and its total count
• sort -n: Optional, sorts the output by PID number for easier readability

Run your script and the reference command above on the same log files, then compare the outputs. If they match exactly, your script is working correctly. If not, debug the discrepancies:

• Did you forget to include one of the log files?
• Is your PID extraction capturing extra characters or missing some valid PIDs?
• Are you counting per file instead of aggregating totals across both inputs?

内容的提问来源于stack exchange，提问作者Subodh