If some records are rejected while writing large amounts of data to Snowflake, the rejected file might not display some of the rejected records even though the statistics of rejected records appear correctly in the session logs.

If you run a mapping configured to create a new target at runtime, you can't write data to a Snowflake target when the table name contains a slash (/) or backslash (\).

If you configure

onError=ABORT_STATEMENT

and

oneBatch=true

load properties in the

Additional Write Runtime Parameters

field in the Target transformation to write data to Snowflake, and the mapping fails, any data written to the target before the failure is not rolled back in the following scenarios:

The mapping contains multiple targets and fails after writing to at least one of them.
The target uses a batch row size.
The target uses an expression for a data-driven operation.

In advanced mode, the following restrictions apply:

The mapping fails when the Snowflake table names contains double quotes. You can include column names with double quotes.
When you specify a table name for a target that you want to create at runtime and you also specify an update override with a different table name, the mapping considers the table name specified during the design time.
You can't call a stored procedure from a custom query from the Source transformation.
If the selected update column in a Target transformation contains the maximum length defined for the column name, the task fails. Ensure that the length of the column name length does not exceed 75 characters.
If you configure
onError=ABORT_STATEMENT
and
oneBatch=true
load properties in the
Additional Write Runtime Parameters
field in the Target transformation to write data to Snowflake, and the mapping fails, any data written to the target before the failure is not rolled back.