How To Migrate Mysql to Postgresql
Hi i spent this week trying to migrate i MySql Database to Postgresql, i tried everything, Mysql Dump compatiple BLA BLA BLA etc etc … until i found this at Github … And it WORKS !!!!!! Here his the Link https://github.com/AnatolyUss/FromMySqlToPostgreSql
FromMySqlToPostgreSql – the database migration tool.
FromMySqlToPostgreSql is a tool, intended to make a process of migration from MySql to PostgreSql as easy and smooth as possible.
Recently I’ve developed NMIG, a database migration app, written in Node.js, and highly inspired by FromMySqlToPostgreSql.
NMIG has a few important functional enhancements.
Due to Node.js asynchronous nature, NMIG executes a part of its tasks in parallel, what makes it up to 3 times faster than FromMySqlToPostgreSql. Visit https://github.com/AnatolyUss/nmig.
However, if you’re not into Node.js, you can still use old and proven FromMySqlToPostgreSql.
- Ease of use – the only thing needed to run this script is the PHP(CLI) interpreter.
- Accuracy of migration the database structure – FromMySqlToPostgreSql converts MySql data types to corresponding PostgreSql data types, creates constraints, indexes, primary and foreign keys exactly as they were before migration.
- Ability to migrate big databases – in order to reduce RAM consumption
FromMySqlToPostgreSql will split each table’s data into several chunks.
The chunk size can be adjusted easily via configuration file.
- Speed of data transfer – in order to migrate data fast
FromMySqlToPostgreSql uses PostgreSQL COPY protocol.
Note: migration of tables, containing “varbinary” or “blob” columns may be considerably slower.
- Ease of monitoring – FromMySqlToPostgreSql will provide detailed output about every step, it takes during the execution.
- Ease of configuration – all the parameters required for migration should be put in one single file, which can be in either “xml” or “json” format.
- Ability to transfer only a data (in case of an existing database).
- PHP (CLI) 5.4 or above
- PDO_MYSQL should be installed and enabled
- PDO_PGSQL should be installed and enabled
- mbstring should be installed and enabled
- register_argc_argv should be enabled (check php.ini).
1. Create a new database.
CREATE DATABASE my_postgresql_database;
2. Download FromMySqlToPostgreSql package and put it on the machine running your PostgreSql.
3. Create configuration file in either “xml” or “json” format and put it on the machine running your PostgreSql.
- sample_config.json and sample_config.xml are examples of configuration file, so you can edit one of them and use for migration.
- Brief description of each configuration parameter will be found at sample_config.json and sample_config.xml
- Make sure, that username, you use in your PostgreSQL connection string, defined as superuser (usually “postgres”)
More info: http://www.postgresql.org/docs/current/static/app-createuser.html
4. Run the script from a terminal.
php /path/to/FromMySqlToPostgreSql/index.php /path/to/FromMySqlToPostgreSql/config[.xml | .json]
5. At the end of migration check log files, if necessary.
Log files will be located in “logs_directory” folder in the root of the package.
Note: “logs_directory” will be created during script execution.
6. In case of any remarks, misunderstandings or errors during migration,
please feel free to email me email@example.com
Current version is 1.4.0
(major version . improvements . bug fixes)
Tested using MySql Community Server (5.6.21) and PostgreSql (9.3).
The entire process of migration 59.6 MB database (49 tables, 570750 rows),
which includes data types mapping, creation of tables, constraints, indexes,
PKs, FKs, migration of data, garbage-collection and analyzing the newly created
PostgreSql database took 3 minutes 6 seconds.
Tested using MySql Community Server (5.6.21) and PostgreSql (9.4).
The entire process of migration 3.1 GB database (56 tables, 8732967 rows),
with the same steps as above, took 2 hours and 10 minutes.
FromMySqlToPostgreSql is available under “GNU GENERAL PUBLIC LICENSE” (v. 3)
Errors/Exceptions are not passed silently.
Any error will be immediately written into the error log file.
Big thanks to all contributors for their valuable remarks!