What is mysqlnothx?

This is a dump file converter that takes MySQL dump files and converts them to PostgreSQL. It is based on a converter I wrote while working at Zalora. Thanks to the Pipes library, it is capable of parsing very large files without running out of memory.

Cool features

• Insert statements are converted to using the COPY tablename FROM stdin command for faster importing
• Indexes and Constraints are moved to the very end of the file, after the data has been inserted, just like a real PostgreSQL dump
• It differentiates between real and "imaginary" foreign key constraints. If the table it's referencing doesn't exist in the dump file, it's placed after a savepoint named imaginary_constraints, allowing you to partially rollback the import if the tables don't exist in the database either.

Installation & Setup

TODO: write this part

Supported MySQL Column Types

• Text (tiny, medium, large)
• Char
• Varchar
• Enum (partial)
• Set (partial)
• Date
• Time
• Datetime
• Timestamp
• Bit
• Blob
• Int (tiny, small, medium, big)
• Double
• Float
• Numeric
• Decimal

In addition to porting these column types to their closest compatible PostgreSQL data type, it has fixes for the following cases:

• CHAR(0) and VARCHAR(0) converts to CHAR(1)
• Default date values of 0000-00-00 for types that include a date (DATE, DATETIME, TIMESTAMP) are converted to CURRENT_TIMESTAMP

TODO

List of features I'd like to add.

• Addressing time zone differences between source and destination (not sure if this is reasonable)
• Fixing out of bounds dates (eg. February 31st)