You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@parquet.apache.org by "Uwe L. Korn (JIRA)" <ji...@apache.org> on 2016/10/14 05:59:20 UTC

[jira] [Created] (PARQUET-749) Schema building ParquetFileWriter

Uwe L. Korn created PARQUET-749:
-----------------------------------

             Summary: Schema building ParquetFileWriter
                 Key: PARQUET-749
                 URL: https://issues.apache.org/jira/browse/PARQUET-749
             Project: Parquet
          Issue Type: New Feature
          Components: parquet-cpp
            Reporter: Uwe L. Korn
            Assignee: Uwe L. Korn


Sometimes you want to write a Parquet file column-by-column without knowing the whole schema upfront. Format-wise this should be possible if you only have a single RowGroup. The user would write column after column and only at the end (when we need to write the metadata), the whole schema can be assembled.

Currently ParquetFileWriter only supports writing files when the schema is known upfront.

Limitations for the start:
 * A single RowGroup is written
 * No nestings



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