Skip to content

Conversation

@igorwwwwwwwwwwwwwwwwwwww

The current behaviour is to load everything into a non-partitioned table. That means that queries will scan the entire table every time.

In order to make queries cheaper, we can use time partitioning.

One nice side-effect of this is that we also get the ability to configure expiration.

Note: I could use some help testing out this patch.

shagamemnon and others added 28 commits July 17, 2019 11:41
small typo fix
According to [docs](https://cloud.google.com/functions/docs/migrating/nodejs-runtimes) and warning emails 

> The Node.js 8 runtime will be deprecated on 2020-06-05

At moment I'm connecting our cloudflare to gcp, and decided to set node to 10 can confirm that everything working
* Added Spectrum schema and bash $SCHEMA variable.

* type fix for Spectrum schema

* add $SCHEMA to automatic install

* fix README.md

* fix capitalization

* Update cloudshell.md

Co-authored-by: Frank Taylor <7483580+shagamemnon@users.noreply.github.com>
@shagamemnon
Copy link
Contributor

Hey @igorwwwwwwwwwwwwwwwwwwww thanks for taking a stab at this. In the past, creating ingestion-time partitioning for logs inserted via a load job has been non-trivial. The solution in your PR has not worked in the past (see https://github.com/cloudflare/cloudflare-gcp/blob/add-partition-v2/logpush-to-bigquery/index.js), so we will need a test before a merge can be considered.

FWIW: to my knowledge, this can only be accomplished in BigQuery using "partition decorators" which are described here: https://cloud.google.com/bigquery/docs/creating-partitioned-tables#creating_an_ingestion-time_partitioned_table_when_loading_data

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

9 participants