gfunc

experienced the same error when upgraded to 0.5.0 release and configed to use grpc port(9100) with sparksql, but things worked when switched to http port (8123) #### jars used -...

> > Please disable compression when you are using gRPC w/ old(before 22.3) ClickHouse and see what will happen. > > ``` > spark.clickhouse.write.compression.codec=none > spark.clickhouse.read.compression.codec=none > ``` the same...

using a fork with support for distributed engine as well, repo [here](https://github.com/gfunc/dbt-clickhouse) my solution to distributed tables was to create the on cluster distributed table with the model name, in...

I started to merge my approach toward distributed table engine. And I want to start a discussion early on which is about the handling of `unique_keys`. My production env has...

> Optimize is by nature an expensive operation, since it most cases it rewrites all of the data in the table. (You can also OPTIMIZE a ReplacingMergeTree table, which is...

Perhaps it would be nice to add a switch for this feature (snapshot data on S3 disk)? For my use case, I need to back up metadata and metadata only.

Hi @genzgd, Thanks for your comment. I think the answer is no. To my understanding, the table materialization (not incremental) now is not affected by the `full-refresh` flag much except...

I will take a look. I think it is a problem with the exchange macro. It should not give the`on cluster` clause for table materialization. Hi @ikeniborn, are you expecting...

I can reproduce the problem now. It seems to be a compatibility issue. model `dimension.dim_twitter_pinned_tweet` already exists on the cluster. But with the latest `dbt-clickhouse` ver 1.4.9 during the creation...

In #206 my solution to this problem is that we provide a detailed message to reflect the error and make `full-refresh` workable in this situation. I am not sure this...