Greg Roberts

It seems absurd that this issue has been open for over 4 years without a solution. We have a mix of plain CF templates and CDK across our estate, and...

We're encountering this issue using iceberg 1.4.3 on Spark 3.3.0 (AWS Glue 4.0) and using `toTable` hasn't fixed the issue.

We've found a workaround in our use case. (Iceberg 1.4.3, Spark 3.3.0 on Glue 4.0). Our previous flow was: ``` TABLE = "glue_catalog.." # set up readStream read_stream = spark.readStream.format(...

@stevenlii > But this way there is no checkpointLocation. How do you manage the offset? As I said, we're using `GlueContext.forEachBatch` which allows you to specify the checkpoint location as...