doc(partner_sdk): Alter Table operation guidelines #71
Conversation
|
This pull request has been linked to 1 task:
|
fivetran-rishabhghosh
requested review from
varundhall,
alsnkline,
shivangisingh172 and
fivetran-dejantucakov
There was a problem hiding this comment.
@fivetran-rishabhghosh Please see suggestions and comments.
Co-authored-by: Dejan Tucakov <141165749+fivetran-dejantucakov@users.noreply.github.com>
|
@fivetran-dejantucakov Updated |
Co-authored-by: Dejan Tucakov <141165749+fivetran-dejantucakov@users.noreply.github.com>
| @@ -125,6 +125,10 @@ This operation should report all columns in the destination table, including Fiv | |||
| - This operation might be requested for a table that does not exist in the destination. In that case, it should NOT fail, simply ignore the request and return `success = true`. | |||
| - `utc_delete_before` has millisecond precision. | |||
|
|
|||
| #### AlterTable | |||
| - This operation should be used for changing primary key columns, adding columns, and changing data types. | |||
| - However, this operation should not drop any columns even if the `AlterTable` request has a table with a different set of columns. Dropping columns could lead to unexpected customer data loss and is against Fivetran's general approach to data movement. | |||
There was a problem hiding this comment.
This doesn't say what they should do. i'd also love to link to a doc that describes our policy/what we usually do - ie not dropping columns/deleting data - but I didn't just find a suitable link.
eg 'this operation should not drop any columns even if the 'AlterTabel' request has a table with a different set of columns. Fivetran's approach to data movement is to leave such columns in a customer's warehouse with no additional data populating to provide the customer control and choice on timing for removal. Dropping columns can lead to unexpected customer data loss.
Links https://fivetran.height.app/T-797932/description
Add guidelines for Alter Table operation