From claude-code-starter
Migrates data between formats (CSV/JSON/Excel/Parquet), databases (MySQL to PostgreSQL/SQLite), and API imports using pandas, pgloader, with backups, dry runs, and integrity checks.
How this skill is triggered — by the user, by Claude, or both
Slash command
/claude-code-starter:migrateThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
帮用户在不同数据格式之间转换,或从一个系统迁移数据到另一个系统。
帮用户在不同数据格式之间转换,或从一个系统迁移数据到另一个系统。
在做任何迁移操作之前,必须先备份源数据:
cp source.csv source.csv.bakpg_dump / mysqldump / sqlite3 db.sqlite .dump > backup.sql根据用户需求,走对应的子路径:
(a) 文件格式转换(CSV/JSON/Excel/Parquet 互转)
file 或 chardetect 确认源文件编码(b) 数据库迁移(MySQL -> PostgreSQL、SQLite -> PostgreSQL 等)
(c) API 数据导入(从外部 API 拉数据存到本地)
在正式迁移前,先用小样本测试:
试运行结果:从前 100 条数据中,成功转换 98 条,2 条有问题(原因:xxx)。 预计全量迁移 N 条记录,耗时约 X 分钟。 确认继续吗?
用户确认后再执行全量迁移。
执行全量迁移,完成后做数据完整性检查:
迁移完成。源数据 N 条,目标 N 条,校验通过。 备份文件在:xxx,确认无误后可删除。
chardetect file.csv 检测实际编码,常见的是 GBK/GB2312 被当成 UTF-8 读取。pandas 用 encoding='gbk' 参数。npx claudepluginhub lightpointventures/claude-code-starterGuides schema and data migrations across engines with tools like Flyway, Liquibase, Alembic, Prisma Migrate. Covers zero-downtime patterns (expand-contract), cross-engine (MySQL→PG), CDC (Debezium). Use for schema changes or DB migrations.
Exports Kongcode data as JSONL for migration to non-SurrealDB systems like Postgres+pgvector, Neo4j, OpenSearch, or custom stores.
Guides Snowflake migrations from Redshift, BigQuery, or on-prem databases including schema conversion, SQL dialect mapping, and data transfer via S3 or other storage.