Join discussions on data engineering best practices, architectures, and optimization strategies within the Databricks Community. Exchange insights and solutions with fellow data engineers.
Hi,I know how filtering a delta table on a partition column is a very powerful time-saving approach, but what if this column appears as a CONCAT in the where-clause?I explain my case: I have a delta table with only one partition column, say called co...
Hi Everybody,I have 20 years of data, 600m rows.I have partitioned them on year and month to generated a files size which seems reasonable.(128Mb)All data is queried using timestamp, as all queries needs to filter on the exact hours.So my requirement...
Hi Guys, thanks for your advices. I found a solution. We upgrade the Databricks Runtime to 12.2 and now the pushdown of the partitionfilter works. The documentation said that 10.4 would be adequate, but obviously it wasn't enough.
When I use SQL code like "create table myTable (column1 string, column2 string) using csv options('delimiter' = ',', 'header' = 'true') location 'pathToCsv'" to create a table from a single CSV file stored in a folder within an Azure Data Lake contai...
Hi @andrew li​, When you specify a path with LOCATION keyword, Spark will consider that to be an EXTERNAL table. So when you dropped the table, you underlying data if any will not be cleared. So in you case, as this is an external table, you folder s...
Here are the simple steps to reproduce it. Note that col "foo" and "bar" are just redundant cols to make sure the dataframe doesn't fit into a single partition. // generate a random df
val rand = new scala.util.Random
val df = (1 to 3000).map(i => (r...
Hi @Jerry Xu​ Thank you for your question! To assist you better, please take a moment to review the answer and let me know if it best fits your needs.Please help us select the best solution by clicking on "Select As Best" if it does.Your feedback wil...
I have a delta table that is partitioned by Year, Date and month. I'm trying to merge data to this on all three partition columns + an extra column (an ID). My merge statement is below:MERGE INTO delta.<path of delta table> oldData
using df newData ...
Isn't the suggested idea only filtering the input dataframe (resulting in a smaller amount of data to match across the whole delta table) rather than prune the delta table for relevant partitions to scan?
Hi,maybe someone can help me i do want to run very narrow query SELECT *
FROM my_table
WHERE snapshot_date IN('2023-01-06', '2023-01-07')
-- part of the physical plan:
-- Location: PreparedDeltaFileIndex [dbfs:/...]
-- PartitionFilters: [cast(snaps...
Hi @Mariusz J​, We haven’t heard from you since the last response from @Werner Stinckens​, and I was checking back to see if his suggestions helped you.Or else, If you have any solution, please share it with the community, as it can be helpful to oth...
We are reading 520GB partitions files from CSV and when we write in a Single CSV using repartition(1) it is taking 25+ hours. please let us know an optimized way to create a single CSV file so that our process could complete within 5 hours.
Hi @mohit kumar suthar​ Hope all is well! Just wanted to check in if you were able to resolve your issue and would you be happy to share the solution or mark an answer as best? Else please let us know if you need more help. We'd love to hear from you...
I'm trying to create delta live table on top of json files placed in azure blob. The json files contains white spaces in column names instead of renaming I tried `columnMapping` table property which let me create the table with spaces but the column ...
I have a large table which contains a date_time column.The table contains 2 generated columns year, and month which are extracted from the date_time values and are used for partitioning.I have the following question.If I run the querySELECT *FROM tab...
Hi @Andrej Znidarsic​ Hope all is well! Just wanted to check in if you were able to resolve your issue and would you be happy to share the solution or mark an answer as best? Else please let us know if you need more help. We'd love to hear from you.T...
Recommendations for performance tuning best practices on DatabricksWe recommend also checking out this article from my colleague @Franco Patano​ on best practices for performance tuning on Databricks.​Performance tuning your workloads is an important...
Is there any way to overwrite a partition in delta table without specifying each and every partition in replace where. For non dated partitions, this is really a mess with delta tables.Most of my DE teams don't want to adopt delta because of these gl...
Hi All,I am trying to Partition By () on Delta file in pyspark language and using command:df.write.format("delta").mode("overwrite").option("overwriteSchema","true").partitionBy("Partition Column").save("Partition file path") -- It doesnt seems to w...
Hey @Harsha kriplani​ Hope you are well. Thank you for posting in here. It is awesome that you found a solution. Would you like to mark Hubert's answer as best? It would be really helpful for the other members too.Cheers!
Situation: we have one partion per date, and it just so happens that each partition ends up (after optimize) as *a single* 128mb file. We partition on date, and zorder on userid, and our query is something like "find max value of column A where useri...
Z-Order will make sure that in case you need to read multiple files, these files are co-located.For a single file this does not matter as a single file is always local to itself.If you are certain that your spark program will only read a single file,...
I am in the process of building my data pipeline, but I am unsure of how to choose which fields in my data I should use for partitioning. What should I be considering when choosing a partitioning strategy?
The important factors deciding partition columns are:Even distribution of data. Choose the column that is commonly or widely accessed or queried. Do not create multiple levels of partition, as you can end up with a large number of small files.
