Hey folks, I wanted to ask about

parse_tokens_limit

, I see the default is set to 3.5 million. I was inserting to a table with ~130 columns with a batch size of 10,000 and it errored out due to going above the token limit. I wnated to know what's the downside of increasing this limit and what's a reasonable upper cap on it? I tried finding it in the docs but didn't see anything on it.

insert into values

with large number of rows is discouraged. not only it consumes a lot of FE memory, but also way inefficient and slow to ingest data into the table. Suggest build the data into files and leverage stream load interface to batch loading into the system.

Thanks Kevin! I will try out stream load API.