Skip to content

Commit 2312043

Browse files
Update docs/t-sql/statements/copy-into-transact-sql.md
Co-authored-by: William Assaf MSFT <wiassaf@microsoft.com>
1 parent ddf28b6 commit 2312043

1 file changed

Lines changed: 3 additions & 1 deletion

File tree

docs/t-sql/statements/copy-into-transact-sql.md

Lines changed: 3 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -430,7 +430,9 @@ WITH (
430430
## FAQ
431431

432432
### What is the performance of the COPY command compared to PolyBase?
433-
The COPY command will have better performance depending on your workload. For best loading performance, consider splitting your input into multiple files when loading compressed CSV.
433+
The COPY command will have better performance depending on your workload.
434+
- Uncompressed files can be split automatically, so there's no need to manually split CSV files.
435+
- Large compressed cannot be split automatically. For best loading performance, consider splitting your input into multiple files when loading compressed CSV.
434436

435437
### What is the file splitting guidance for the COPY command loading compressed CSV files?
436438
Guidance on the number of files is outlined in the table below. Once the recommended number of files are reached, you will have better performance the larger the files. The number of files is determined by number of compute nodes multiplied by 60. For example, at 6000DWU we have 12 compute nodes and 12*60 = 720 partitions. For a simple file splitting experience, refer to [How to maximize COPY load throughput with file splits](https://techcommunity.microsoft.com/t5/azure-synapse-analytics/how-to-maximize-copy-load-throughput-with-file-splits/ba-p/1314474).

0 commit comments

Comments
 (0)