因为我的数据库更新是通过 Liquibase 执行的，所以每次发版的时候都会新增 liquibase 记录，如果没有一个自动检查的 pipeline， 很容易在发版的时候因为有重复的 changeset id 这一小失误导致pipeline 被中断

于是，我通过在 Github workflow 中添加一个每次有新提交时就扫描所有文件内是否有重复 changeset id 出现的 workflow 来规避这个失误

workflow 文件: .github/workflows/check-duplications.yml

name: Check Changeset Duplications

on:
  push:
    paths:
      - '**.sql'
  
jobs:
  changeset-checking:
    runs-on: ubuntu-latest
    steps:
    - name: Checkout repository
      uses: actions/checkout@v4

    - name: Checking duplications
      run: bash liquibase/scripts/check_changeset_duplications.sh

shell 脚本: check_changeset_duplications.sh

#!/bin/bash

extract_ids() {
  local file=$1
  local changeset_ids=()
  local file_lines=()

  while IFS= read -r line; do
    id=$(echo $line | perl -lne 'print $1 if m{^\d+:--\s*changeset (.*?)(?= runOnChange| endDelimiter|;|$)}')
    line_number=$(echo $line | perl -lne 'print $1 if m{^(\d+)}')
    if [[ -n "$id" && -n "$line_number" ]]; then
      if echo "$id" | grep -q " "; then
        id=$(echo $id | tr ' ' '_')

      fi
      changeset_ids+=("$id")
      file_lines+=("$file:$line_number")
    fi
  done < <(perl -lne 'print "$.:$_" if /--\s*changeset .*?(?= runOnChange| endDelimiter|;|$)/' $file)

  echo "${changeset_ids[*]}|${file_lines[*]}"
}

check_duplicates() {
  IFS=' ' read -r -a changeset_ids <<<"$1"
  IFS=' ' read -r -a file_lines <<<"$2"
  local duplicates=()
  local checked_ids=()

  for index in ${!changeset_ids[@]}; do
    id=${changeset_ids[$index]}
    if [[ " ${checked_ids[@]} " =~ " ${id} " ]]; then
      continue
    fi
    checked_ids+=("$id")

    count=$(printf '%s\n' "${changeset_ids[@]}" | grep -c "$id")
    if [ $count -gt 1 ]; then
      first_occurrence_found=false
      for second_index in ${!changeset_ids[@]}; do
        if [ "${changeset_ids[$second_index]}" = "$id" ]; then
          if [ "$first_occurrence_found" = true ]; then
            # echo "Duplicate changeset ID: $id found at ${file_lines[$second_index]}"
            echo -e "Duplicate changeset ID: \e[0;35m$id\e[0m found at file:file_line (\e[0;35m${file_lines[$second_index]}\e[0m) ❌❌❌"
          else
            first_occurrence_found=true
          fi
        fi
      done
    fi
  done
}

total_duplicates=0
for file in $(find . -mindepth 1 -type d -name 'Testing' -prune -o -type f -name "*.sql" -print); do
    echo -e "------------------------------------------Begin: $file------------------------------------------"
    echo -e "---- Checking file $file..."
    result=$(extract_ids $file)

    changeset_ids=(${result%|*})
    file_lines=(${result#*|})

    echo -e "-------- Total ${#changeset_ids[*]} changeset id in $file "
    echo -e "-------- Total ${#file_lines[*]} file lines in $file "
    tmpfile=$(mktemp)
    check_duplicates "${changeset_ids[*]}" "${file_lines[*]}" >"$tmpfile"
    total_duplicates_in_file=$(wc -l <"$tmpfile")
    total_duplicates=$((total_duplicates + total_duplicates_in_file))
    if [ $(wc -c <"$tmpfile") -gt 0 ]; then
      cat "$tmpfile"
    fi
    rm "$tmpfile"
    echo -e "------------------------------------------End: $file------------------------------------------\n"
done
echo -e "------------------------------------------All done------------------------------------------\n"

if [ $total_duplicates -gt 0 ]; then
  echo "---- Total $total_duplicates duplicates be found."
  exit 1
else
  echo "-------- Congratuations:"
  echo "---- No duplicate changeset IDs found. All changeset IDs in the files are unique."
  echo "---- Good to go."
fi

检查效果：
1. 有重复项时，打印：重复的changeset id, 所在文件名，所在文件中的行号
2. 没有重复项时，打印：Good to go.