Several Problems During sqlserver2postgresql migration

1. PostgreSQL cross-platform migration tool migration
Toolkit User Guide: http://www.enterprisedb.com/docs/en/8.4/mtkguide/Table%20of%20Contents.htm#TopOfPage

2. It may beProgramThe migration tool considers the jtds driver connection string as Oracle to be connected:

Java-dprop = toolkit. properties-jar lib/edb-migrationtoolkit.jar-sourcedbtype sqlserver-targetdbtype postgresql-alltables-Targetschema
Public DBO

3. During the migration using the above tools, the type matching problem that can impede the migration of the entire table is discovered, the bit type of MSSQL is created as a Boolean field during table creation in PostgreSQL, but an error is reported during data loading: the BIT data is converted into a boolean data by legal means. I was just preparing to write a Python script to migrate these tables. Later I found that the problem was relatively simple. I directly converted the bit type in the source table to the smartint type and then migrated it, prepare to go back to PostgreSQL.

4. After the above problem is mentioned, 3. After the data migration is completed, the configuration file of the Web application data source is modified, the Web application is started, and the Read and Write tests of tables of the type changed earlier are performed, normal availability was found. Previously, the bit type in the hibernate ORM configuration file automatically generated by the bit type in sqlserver was Java. lang. byte and byte are 8-bit bytes, and smallint is exactly 8-bit, which is compatible and available. It is an accidental luck event.

5. Let's talk about another problem. When the tool data was migrated, the system often reminded me of the UTF-8 encoding problem. I thought that all the string data will be automatically converted to the UTF-8 data during the migration, before the web application is GBK encoding, so I guess it is not the completion of the migration has to do the replica set all converted to UTF-8, but the actual discovery is not like this, I used python to query the string with the U character header in sqlserver. It seems that the U character header is no longer included in PostgreSQL.