Compared to the native spark, the shuffle write data of firestorm is always smaller

[xunxunmimi5577]

Due to company's security restrictions,I cannot upload images. Let me describe the phenomenon simply , for the same stage of the same SQL, the value of Shuffle Write Size/Records are 16 B/1 and 58 B/1 respectively? The former is the result from firestorm

[jerqi]

Firestorm will compress data. So it's normal phenomeonon.

[xunxunmimi5577]

ok，I see，thank you!