topic Re: How to download the results in batches in Get Started Discussions

How to download the results in batches

Lumoura — Tue, 28 Jan 2025 21:43:51 GMT

Hello, how are you?

I`m trying to download some of my results on databricks and the sheets is around 300mb, unfortunately my google sheets is not open files that has more then 100mb. Is that any chance that i could download the results in batches to join after manually? (The results is more then 100k lines)

Re: How to download the results in batches

szymon_dybczak — Wed, 29 Jan 2025 10:07:59 GMT

Hi @Lumoura ,

So maybe just save the result as a csv to Volume and then use pyspark to split this file into smaller parts, i.e using repartition:

df = df.repartition(num_partitions)

Re: How to download the results in batches

NandiniN — Sat, 01 Feb 2025 06:24:23 GMT

Hey,

Thinking of more alternates to repartition:

1- Use the limit and offset options in your SQL queries to export data in manageable chunks. For example, if you have a table with 100,000 rows and you want to export 10,000 rows at a time, you can use the following queries:

SELECT * FROM your_table LIMIT 10000 OFFSET 0; SELECT * FROM your_table LIMIT 10000 OFFSET 10000; SELECT * FROM your_table LIMIT 10000 OFFSET 20000; ...

(Adjust the LIMIT and OFFSET values based on the size of your data and the number of rows you want to export in each batch)

2- Save the results of each query to a separate CSV file. You can use the dbutils.fs module to save the results to DBFS (Databricks File System) and then download them to your local machine.

3 - My personal fav is splitting, when working with heap dumps, you could use that too once you have the large file, split them up.

split -b 100M large.tar.gz large.tar.gz.part