I have a few processes that yield a partitioned table in BigQuery and, when the datasource has issues I end up with, say 100 partitions of a table that are useless and I need to delete them.
Does anyone know how to achieve that via the BQ UI?
I've tried using the DML with no result... apparently it only deletes the content within the partitions, not the partitions.
Any particular reasons to physically delete the partitions instead of clear content? Alternatively, you can try
bq rm -f -t [PROJECT_ID]:[DATASET].[TABLE]$[20170101]
Related
In our BQ export schema, we have one table for each day as per the screenshot below.
I want to copy the tables before a certain date (2021-feb-07). I know how to copy one day at a time via the UI, but is there not a way to use the cloud console to write a code for copying the selected date range, all at once? Or maybe an sql command directly from a query window?
I think you should transform your sharding tables into a partitioned table. So you can handled your tables with just a single query. As mention in the official documentation, partitioned tables perform better.
To make the conversion, you can just execute the following commands in the console.
bq partition \
--time_partitioning_type=DAY \
--time_partitioning_expiration 259200 \
mydataset.sourcetable_ \
mydataset.mytable_partitioned
This will make your sharded tables sourcetable_(xxx) into a single partitioned table mytable_partitioned which can be query with just a single query trough your entire set of data entries.
SELECT
*
FROM
`myprojectid.mydataset.mytable_partitioned`
WHERE
_PARTITIONTIME BETWEEN TIMESTAMP('2022-01-01') AND TIMESTAMP('2022-01-03')
For more details about the conversion commands you can check this link. Also, I recommend to check the links about querying partionated tables and partiotioned tables for more details.
I have a dataset, which consists of ~600 tables. I want to create a new dataset with no tables and same name using CLI.
At the moment I'm iterating through all of the tables and dropping them with "bq rm" 1 by one, but it takes ~20min. Maybe I can simply drop dataset without removing the tables first?
Use the -r flag. For example:
bq rm -rf dataset_name
The -f flag means "force", so the command won't prompt for confirmation.
You can just simply use contextual menu in BigQuery UI to delete dataset
When you try and dataset has tables in it - you will see below form
Just type your dataset_name to confirm deletion - and you are done!
Is there any number of partitions we would expect this command
MSCK REPAIR TABLE tablename;
to fail on?
I have a system that currently has over 27k partitions and the schema changes for the Athena table we drop the table, recreate the table with say the new column(s) tacked to the end and then run
MSCK REPAIR TABLE tablename;
We had no luck with this command doing any work what so every after we let it run for 5 hours. Not a single partition was added. Wondering if anyone has information about a partition limit we may have hit but can't find documented anywhere.
MSCK REPAIR TABLE is an extremely inefficient command. I really wish the documentation didn't encourage people to use it.
What to do instead depends on a number of things that are unique to your situation.
In the general case I would recommend writing a script that performed S3 listings and constructed a list of partitions with their locations, and used the Glue API BatchCreatePartition to add the partitions to your table.
When your S3 location contains lots of files, like it sounds yours does, I would either use S3 Inventory to avoid listing everything, or list objects with a delimiter of / so that I could list only the directory/partition structure part of the bucket and skip listing all files. 27K partitions can be listed fairly quickly if you avoid listing everything.
Glue's BatchCreatePartitions is a bit annoying to use since you have to specify all columns, the serde, and everything for each partition, but it's faster than running ALTER TABLE … ADD PARTION … and waiting for query execution to finish – and ridiculously faster than MSCK REPAIR TABLE ….
When it comes to adding new partitions to an existing table you should also never use MSCK REPAIR TABLE, for mostly the same reasons. Almost always when you add new partitions to a table you know the location of the new partitions, and ALTER TABLE … ADD PARTION … or Glue's BatchCreatePartitions can be used directly with no scripting necessary.
If the process that adds new data is separate from the process that adds new partitions, I would recommend setting up S3 notifications to an SQS queue and periodically reading the messages, aggregating the locations of new files and constructing the list of new partitions from that.
Is there a way to copy a date-sharded table to another dataset via the bq utility?
My current solution is generating a bash script to copy each day one-by-one and splitting the work, but more efficient would be to do everything in parallel:
#!/bin/sh
bq cp old_dataset.table_20140101 new_dataset_20140101
..
bq cp old_dataset.table_20171001 new_dataset_20171001
You can specify multiple source tables but only a single destination table (refer to this question), so this may not work for you. However, if your data is date-partitioned (instead of sharded) then you can copy the table in one command.
I recommend you convert the sharded table into a date-partitioned table which will be effectively copying all the sharded tables to a new table. You can do this with the following command:
bq partition old_dataset.table_ new_dataset.partitioned
I'm not a developer so please bear with me on this. I wasn't able to follow the PHP-based answer at Google BigQuery - Automating a Cron Job, so I don't know if that's even the same thing as what I'm looking for.
Anyway, I use Google Cloud to store data, and several times throughout the day data is uploaded into CSVs there. I use BigQuery to run jobs to populate BigQuery tables with the data there.
Because of reasons beyond my control, the CSVs have duplicate data. So what I want to do is basically create a daily ETL to append all new data to the existing tables, perhaps running at 1 am every day:
Identify new files that have not been added (something like date = today - 1)
Run a job on all the CSVs from step 1 to convert them to a temporary BigQuery table
De-dupe the BigQuery table via SQL (I can do this in a variety of ways)
Insert the de-duped temp table into the BigQuery table.
Delete the temp table
So basically I'm stuck at square 1 - I don't know how to do any of this in an automated fashion. I know BigQuery has an API, and there's some documentation on cron jobs, and there's something called Cloud Dataflow, but before going down those rabbit holes I was hoping someone else may have had experience with this and could give me some hints. Like I said, I'm not a developer so if there's a more simplistic way to accomplish this that would be easier for me to run with.
Thanks for any help anyone can provide!
There are a few ways to solve this, but I'd recommend something like this:
Create a templated Dataflow pipeline to read from GCS (source) and write append to BigQuery (sink).
Your pipeline can remove duplicates directly itself. See here and here.
Create a cloud function to monitor your GCS bucket.
When a new file arrives, your cloud function is triggered automatically, which calls your Dataflow pipeline to start reading the new file, deduping it and writing the results to BigQuery.
So no offense to Graham Polley but I ended up using a different approach. Thanks to these pages (and a TON of random Batch file Google searching and trial and error):
how to get yesterday's date in a batch file
https://cloud.google.com/bigquery/bq-command-line-tool
cscript //nologo C:\Desktop\yester.vbs > C:\Desktop\tempvar.txt &&
set /p zvar =< C:\Desktop\tempvar.txt &&
del C:\Desktop\tempvar.txt &&
bq load
--skip_leading_rows=1
data.data_%%zvar:~0,4%%%%zvar:~4,2%%%%zvar:~6,2%%_1
gs://mybucket/data/%%zvar:~0,4%%-%%zvar:~4,2%%-%%zvar:~6,2%%*.csv.gz
Timestamp:TIMESTAMP,TransactionID:STRING &&
bq query --destination_table=data.data_%%zvar:~0,4%%%%zvar:~4,2%%%%zvar:~6,2%%2 "SELECT * FROM data.data%%zvar:~0,4%%%%zvar:~4,2%%%%zvar:~6,2%%_1 group by 1,2" &&
bq cp -a data.data_%%zvar:~0,4%%%%zvar:~4,2%%%%zvar:~6,2%%_2 data.data &&
bq rm -f data.data_%%zvar:~0,4%%%%zvar:~4,2%%%%zvar:~6,2%%_1 &&
bq rm -f data.data_%%zvar:~0,4%%%%zvar:~4,2%%%%zvar:~6,2%%_2
A VB script called yester.vbs prints out yesterday's date in YYYYMMDD format. This is saved as a variable which is used to search for yesterday's data files in GCS and output to a table, from which a de-duped (via grouping by all columns) table is created. This is then appended to the main table, and the two intermediate tables are deleted.
The double percent signs are shown because it's saved as .CMD file and run through Windows Task Scheduler.