Hi Nathan,
I can share my recent experience for using changlog file generated from Oracle and used it for HANA database, the following adjustments I had to make to use the same changlog successfullly, in the end they seem minor changes and worth time spent as the changes applied to HANA went as expected.
Tables :
- removed default value "sysdate" from the column definition
- changed data type: binary_double to double
- changed data type: Long raw to blob
DATA file :
- No change required, all data loads successfully including CLOB fields
Sequences:
Sequence creation worked after removing clauses : cache and order
Primary key constraints:
Removed the following pattern as HANA does not create physical index for primary key
- <createIndex.*?/createIndex>
- forIndexName.*?PK1"
Unique key constraints:
Removed the following pattern as HANA does not create physical index for unique key
- <createIndex.*?/createIndex>
- forIndexName.*?UK1"
Foreign Key constraints:
- corrected datatype from CHAR to VARCHAR (HANA does not support constraints on CHAR columns)
Indexes
- corrected the following line as HANA does not support "SYS_OP_C2C"
<column computed="true" name="SYS_OP_C2C(<COLUMN_NAME>)"/>
In addition to above, we do have few procedures to be created on HANA. Since they are less, i have created xml files to be loaded into HANA, however, did not succeed so far as "dbms=hana" does not seem to work well with HANA and procedure is not able to be created.
Let me know if you need more information.
Thanks & regards,
Viren