Selective overwrite on Partition and Liquid clustered tables

RGSLCA
New Contributor II

Friday - last edited Friday

Hi,

I have created 2 identical tables but one is partitioned and the one is a Liquid Clustered with Auto Clustering.

I inserted 30M rows x 2 (60M) for two dates , date 1 = 2026-06-01 and date = 2026-06-02 , then I overwrite the date 2026-06-02 with a selective overwrite statement 

REPLACE USING (BED)

For partition table , history shows ,

operationParameters 
 

object

  • mode: 
    "Overwrite"
  • statsOnLoad: 
    "false"
  • partitionBy: 
    "[]"
operationMetrics 
 object
  • numFiles: 
    "43"
  • numRemovedFiles: 
    "43"
  • numRemovedBytes: 
    "6090798096"
  • numDeletionVectorsRemoved: 
    "0"
  • numOutputRows: 
    "30000000"
  • numOutputBytes: 
    "6101142740"

For Liquid Clustered table 

operationParameters 
 object
  • partitionBy: 
    "[]"
  • clusterBy: 
    "[]"
  • mode: 
    "Overwrite"
  • clusteringOnWriteStatus: 
    null
  • statsOnLoad: 
    "false"
  • replaceUsingCols: 
    "(BED)"
operationMetrics 
 object
  • numFiles: 
    "43"
  • numRemovedFiles: 
    "43"
  • numRemovedBytes: 
    "5779903887"
  • numCopiedRows: 
    "0"
  • numDeletionVectorsAdded: 
    "0"
  • numDeletedRows: 
    "30000000"
  • numDeletionVectorsRemoved: 
    "0"
  • numAddedChangeFiles: 
    "0"
  • numOutputRows: 
    "30000000"
  • numOutputBytes: 
    "5779715681"
     

Its overwriting 43 files (Total 86) ?, is this optimal ?

Is there a way to improve performance by reducing the # of files ?

I am using 

INSERT INTO <target>  REPLACE USING (col) 

SELECT <cols> FROM <table> 

 
Labels:
0 Kudos