Split out from the original single-file devops-learning.md. See also
04-shell-pipes-and-text-toolkit.md and 07-regex-from-zero-to-advanced.md.


Part 5 — grep, In Depth

grep PATTERN [files] prints lines matching a pattern. It’s the tool I used most on your codebase.

The flags that matter (ranked by how often you’ll use them)

grep -r  "pattern" src/     # recursive through a directory
grep -n  "pattern" file     # show line numbers        → clickable file:line
grep -i  "pattern" file     # case-insensitive
grep -l  "pattern" src/     # only FILENAMES that match (not the lines)
grep -v  "pattern"          # INVERT — lines that DON'T match (a filter)
grep -c  "pattern" file     # count matching lines
grep -E  "a|b"              # extended regex (see Part 7)
grep -F  "literal.string"   # fixed string, NO regex — fast, safe for special chars
grep -w  "id"               # whole word only (won't match "pillarId")
grep -A5 "pattern"          # show 5 lines After each match
grep -B2 "pattern"          # 2 lines Before
grep -C3 "pattern"          # 3 lines of Context both sides
grep --include="*.ts" -r    # recursive, but only into matching filenames

Real commands from this project, dissected

# "Is faithBackground actually USED anywhere in the backend?"
grep -rn "faithBackground\|equipmentAccess" backend/src --include="*.ts" \
  | grep -v node_modules | grep -v ".spec" | head -20
  • -rn → recursive + line numbers.
  • "faithBackground\|equipmentAccess" → OR. In basic grep, alternation is \| (escaped). With -E it’s a clean |.
  • --include="*.ts" → skip .js builds, .json, etc.
  • | grep -v node_modules | grep -v ".spec" → chained negative filters: drop dependency and test noise. Piping grep into grep -v is the everyday way to subtract.
  • | head -20 → cap the output.
  • The result (only 4 hits, all in onboarding) was the finding: the fields were collected but never used in quest generation. Absence of matches is evidence too.
# "Which onboarding screens mention faith?" — filenames only:
grep -rn "faithBackground\|equipmentAccess\|faith" mobile/app --include="*.tsx" -l
 
# "Show me the seed script config inside package.json, with context":
grep -A5 '"scripts"\|"prisma"' package.json | head -20
#     ^ -A5 prints the 5 lines AFTER each match — perfect for JSON blocks
 
# "Find the message-array boundaries in seed.ts before surgical deletion":
grep -n "LEGACY_MESSAGES_REMOVED\|Seed complete\|existingMessages\|^main()\|^}" prisma/seed.ts
#        ^ anchors: ^main() = lines STARTING with main(), ^} = closing braces at column 0
#        This gave exact line numbers (335, 388) to feed into sed.
 
# case-insensitive filename hunting:
ls -la backend | grep -i env         # found .env and .env.prod.example

grep pro-habits

  • Start broad (grep -rl), then narrow (grep -n in the specific file). Cheap reconnaissance before reading files.
  • grep -F when searching for strings full of dots/brackets (grep -F "toISOString()") — no escaping needed.
  • Know that modern alternatives exist: rg (ripgrep) is grep -r but faster and skips .gitignored files by default. Same flags mostly.


The general version of this

  • Linux (devops course) — note that this file has no general counterpart yet; grep/find/regex/sed depth currently lives only here. Promoting it to devops/01-linux/ is a known gap, tracked in the projects README.
  • sed & awk — the one piece that does exist generally

↑ All projects and the domains they exercise

↑ Back to arete — Learning Notes