Uwe L. Korn created PARQUET-591:
-----------------------------------

             Summary: Page size estimation during writes
                 Key: PARQUET-591
                 URL: https://issues.apache.org/jira/browse/PARQUET-591
             Project: Parquet
          Issue Type: Improvement
          Components: parquet-cpp
            Reporter: Uwe L. Korn


Currently we start a new page after 1000 rows. Instead we should split pages by 
their size. Therefore implement a size estimation algorithm and split pages 
based on its results.

See also in parquet-column/../column/impl/ColumnWriteStoreV2.java:sizeCheck for 
the Java implementation.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to