[SUPPORT]Order changed while using sanpshot query

[homchen-x]

is it normal for hudi that the order of the data will be changed? How can I keep the same order of the data?

[codope]

I guess you are talking about order of the columns? Typically, Hudi will append some metadata fields in the beginning.
If it's not about column reordering, is it about some record versions whose order changed?
Could you please share a reproducible example? Also, is it happening only for snapshot queries?

[liujinhui1994]

Hudi guarantees the order.
Generally, if you use spark to consume kafka to write to hudi, you need to ensure that kafka is in order in the same partition