topic Re: Don't want checkpoint in delta in Data Engineering

Don't want checkpoint in delta

User16826994223 — Tue, 22 Jun 2021 11:53:45 GMT

Suppose I am not interested in checkpoints, how can I disable Checkpoints write in delta

Re: Don't want checkpoint in delta

brickster_2018 — Tue, 22 Jun 2021 22:57:23 GMT

Checkpoint creation in Delta is not user-controllable features/options. Although it's possible to delay the checkpoint file creation, this could have an impact on the performance of the Delta table. By default a checkpoint file creation is triggered for every 10 commits happening on the Delta table.

Re: Don't want checkpoint in delta

sajith_appukutt — Wed, 23 Jun 2021 00:13:57 GMT

Writing statistics in a checkpoint has a cost which is visible usually only for very large tables. However it is worth mentioning that, this statistics would be very useful for data skipping which speeds up subsequent operations.

In Databricks Runtime 7.2 and below, column-level statistics are stored in Delta Lake checkpoints as a JSON column. In Databricks Runtime 7.3 LTS and above, column-level statistics are stored as a struct (struct format makes Delta Lake reads much faster)

There are two flags that control column-level statistics in checkpoints

delta.checkpoint.writeStatsAsJson & delta.checkpoint.writeStatsAsStruct If both table properties are false, no statistics are collected or written - and readers won't be able to perform data skipping.

For more details on tradeoffs with statistics and checkpoints, see here