taler-docs

Documentation for GNU Taler components, APIs and protocols
Log | Files | Refs | README | LICENSE

commit b4fd7989a0a1fd7ce8d74e7e533e0032afc4e94f
parent 1aa71208ff871d41349eb456a901cfc7ed5bd383
Author: Antoine A <>
Date:   Thu, 20 Feb 2025 18:06:31 +0100

DD58 update migration solution

Diffstat:
Mdesign-documents/058-ebics-tx-unique-id.rst | 14++------------
1 file changed, 2 insertions(+), 12 deletions(-)

diff --git a/design-documents/058-ebics-tx-unique-id.rst b/design-documents/058-ebics-tx-unique-id.rst @@ -43,14 +43,4 @@ We should then be able to store all transaction and support UETR everywhere in t Migration ========= -There is two way to migrate the database schema: - -SQL only --------- - -A solution using only SQL would be to rename the bank_id column to compat_id and we would check the compat_id using the current id selection logic. I would prefer not to keep another column just for compatibility as checking three column is already complicated enough. - -SQL and code ------------- - -We can use some non SQL logic during the migration to move the current bank_id to the appropriate column based on the configured dialect. I am not sure how to do that cleanly. -\ No newline at end of file +Before running the SQL migration script we can store the current dialect in a Postgres session variable that will then be used to move the current bank_id into the appropriate column. +\ No newline at end of file