azure-devops - Azure Pipelines - CI/CD：如何针对 DB 执行文件夹中的所有 .sql 文件

Question

我已将所有 SQL 文件签入 Azure devops 上的存储库。我有命名约定，可以让我知道其他 .sql 文件使用了哪些 .sql 文件（例如，文件创建了一个存储过程使用的视图）。我想强制使用 repo 来跟踪代码更改，并且不希望使用 dacpac 文件。我希望每个函数/视图/存储过程都有自己的文件。

我的问题是，我将如何针对来自天蓝色管道的数据库执行与“ ..\Functions\BASE_*.sql ”匹配的所有 .sql 文件？我尝试了以下方法，但不支持匹配多个文件。有没有更好的选择呢？我需要编写一个循环脚本并自己做吗？

# pipeline

trigger:
- master

pool:
  vmImage: 'windows-latest'

steps:
- task: SqlDacpacDeploymentOnMachineGroup@0
  inputs:
    TaskType: 'sqlQuery'
    SqlFile: '$(System.DefaultWorkingDirectory)\Functions\BASE_*.sql'
    ServerName: '$(SQL_ServerName).database.windows.net'
    DatabaseName: '$(SQL_DatabaseName)'
    AuthScheme: 'sqlServerAuthentication'
    SqlUsername: '$(SQL_UserName)'
    SqlPassword: '$(SQL_Password)'

我得到的错误是：

Starting: SqlDacpacDeploymentOnMachineGroup
==============================================================================
Task         : SQL Server database deploy
Description  : Deploy a SQL Server database using DACPAC or SQL scripts
Version      : 0.3.23
Author       : Microsoft Corporation
Help         : https://docs.microsoft.com/azure/devops/pipelines/tasks/deploy/sql-dacpac-deployment-on-machine-group
==============================================================================
##[error]Found more than one file to deploy with search pattern d:\a\1\s\Functions\BASE_*.sql. There can be only one.
Finishing: SqlDacpacDeploymentOnMachineGroup

标签： azure-devopsazure-sql-databasesql-function

Answer 1

经过一天的研究和试用，我能想到的最好办法是在存储库中保持文件分开，然后在 CI/CD 管道中将多个文件组合在一起，然后再针对数据库运行它。

我创建了一个模板来将匹配的文件组合到暂存目录中的单个文件中，发布它以调试管道，然后针对 SQL 服务器执行它。

模板是：

# Template for executing all SQL files matching a string search

parameters:
- name: path #$path = "$(System.DefaultWorkingDirectory)\Functions"
  type: string
- name: match #$match = "BASE_*.sql"
  type: string
- name: outPath #$outPath = "$(System.DefaultWorkingDirectory)\Functions"
  type: string
- name: outName #$outName = "BASE.sql"
  type: string

steps:
- task: PowerShell@2
  inputs:
    targetType: 'inline'
    script: |
      echo Source Files:
      Get-ChildItem ${{parameters.path}} -include ${{parameters.match}} -rec 
  displayName: 'Files to process: ${{parameters.match}}'
- task: PowerShell@2
  inputs:
    targetType: 'inline'
    script: |
      echo Creating: ${{parameters.outPath}}\${{parameters.outName}}
      Get-ChildItem ${{parameters.path}} -include ${{parameters.match}} -rec | ForEach-Object {gc $_; ""} | out-file ${{parameters.outPath}}\${{parameters.outName}}
  displayName: 'Combine: ${{parameters.outName}}'
- task: PublishPipelineArtifact@1
  inputs:
   targetPath: '${{parameters.outPath}}\${{parameters.outName}}'
   artifact: '${{parameters.outName}}'
   publishLocation: 'pipeline'
  displayName: 'Publish: ${{parameters.outName}}'
- task: SqlDacpacDeploymentOnMachineGroup@0
  inputs:
    TaskType: 'sqlQuery'
    SqlFile: '${{parameters.outPath}}\${{parameters.outName}}'
    ServerName: '$(SQL_ServerName).database.windows.net'
    DatabaseName: '$(SQL_DatabaseName)'
    AuthScheme: 'sqlServerAuthentication'
    SqlUsername: '$(SQL_UserName)'
    SqlPassword: '$(SQL_Password)'
  displayName: 'Create or Alter: ${{parameters.outName}}'
- task: PowerShell@2
  inputs:
    targetType: 'inline'
    script: Remove-Item ${{parameters.path}}\${{parameters.match}} -Recurse
  displayName: 'Delete Files: ${{parameters.match}}'

然后主管道使用不同的搜索字符串调用模板。

trigger:
- master

pool:
  vmImage: 'windows-latest'

steps:
- task: PowerShell@2
  inputs:
    targetType: 'inline'
    script: MKDIR "$(System.DefaultWorkingDirectory)\\Combined\\Functions"
  displayName: 'Create Output Folder'
- template: azTemplate/CombineAndRunSQLFiles.yml # Functions: UTIL
  parameters:
    path: "$(System.DefaultWorkingDirectory)\\Functions"
    match: "UTIL_*.sql"
    outPath: "$(System.DefaultWorkingDirectory)\\Combined\\Functions"
    outName: "UTIL.sql"
- template: azTemplate/CombineAndRunSQLFiles.yml # Functions: BASE
  parameters:
    path: "$(System.DefaultWorkingDirectory)\\Functions"
    match: "BASE_*.sql"
    outPath: "$(System.DefaultWorkingDirectory)\\Combined\\Functions"
    outName: "BASE.sql"

结果：

Pool: Azure Pipelines
Image: windows-latest
Agent: Hosted Agent
Started: Today at 9:55 AM
Duration: 1m 6s

Job preparation parameters
5 artifacts produced
Job live console data:
Finishing: Job