[ 
https://issues.apache.org/jira/browse/PARQUET-211?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14394047#comment-14394047
 ] 

Ryan Blue commented on PARQUET-211:
-----------------------------------

I've been going through the command-line semver tool's reports for changes 
between 1.5.0 and the current master. There are lots of breaking changes in 
{{parquet.column}}, but I think that entire package is internal or SPI rather 
than API. I've flagged the following to fix:
* {{ParquetInputSplit#getExtraMetadata}}, {{#getFileSchema}}, 
{{#getRequestedSchema}}, {{#getBlocks}}, and {{#getReadSupportMetadata}} were 
removed and should be added back (This is now documented internal, but was part 
of the API and had external users).
* {{ParquetWriter}} constructors may have an incompatible change

And someone needs to look into this one:
* {{ParquetScroogeScheme#sink}} and {{#isSink}} were removed. 
({{ScroogeStructConverter}} had removals, but I consider it internal)

Lastly, there are quite a few incompatible changes to {{parquet.metadata}} (see 
below). Is this public? It seems like it is part of the API because the 
metadata is exposed. Fixing it will be annoying because {{ColumnPath}} was 
removed entirely.

{code:title=parquet.metadata changes}
Class parquet.hadoop.metadata.Canonicalizer
 Removed Class , access public super synchronized
Class parquet.hadoop.metadata.ColumnChunkMetaData
 Added Method getPath, desc ()Lparquet/common/schema/ColumnPath;, access public
 Removed Method get, sig 
(Lparquet/hadoop/metadata/ColumnPath;Lparquet/schema/PrimitiveType$PrimitiveTypeName;Lparquet/hadoop/metadata/CompressionCodecName;Ljava/util/Set<Lparquet/column/Encoding;>;Lparquet/column/statistics/Statistics;JJJJJ)Lparquet/hadoop/metadata/ColumnChunkMetaData;,
 desc 
(Lparquet/hadoop/metadata/ColumnPath;Lparquet/schema/PrimitiveType$PrimitiveTypeName;Lparquet/hadoop/metadata/CompressionCodecName;Ljava/util/Set;Lparquet/column/statistics/Statistics;JJJJJ)Lparquet/hadoop/metadata/ColumnChunkMetaData;,
 access public static
 Added Method get, sig 
(Lparquet/common/schema/ColumnPath;Lparquet/schema/PrimitiveType$PrimitiveTypeName;Lparquet/hadoop/metadata/CompressionCodecName;Ljava/util/Set<Lparquet/column/Encoding;>;JJJJJ)Lparquet/hadoop/metadata/ColumnChunkMetaData;,
 desc 
(Lparquet/common/schema/ColumnPath;Lparquet/schema/PrimitiveType$PrimitiveTypeName;Lparquet/hadoop/metadata/CompressionCodecName;Ljava/util/Set;JJJJJ)Lparquet/hadoop/metadata/ColumnChunkMetaData;,
 access public static
 Added Method get, sig 
(Lparquet/common/schema/ColumnPath;Lparquet/schema/PrimitiveType$PrimitiveTypeName;Lparquet/hadoop/metadata/CompressionCodecName;Ljava/util/Set<Lparquet/column/Encoding;>;Lparquet/column/statistics/Statistics;JJJJJ)Lparquet/hadoop/metadata/ColumnChunkMetaData;,
 desc 
(Lparquet/common/schema/ColumnPath;Lparquet/schema/PrimitiveType$PrimitiveTypeName;Lparquet/hadoop/metadata/CompressionCodecName;Ljava/util/Set;Lparquet/column/statistics/Statistics;JJJJJ)Lparquet/hadoop/metadata/ColumnChunkMetaData;,
 access public static
 Removed Method getPath, desc ()Lparquet/hadoop/metadata/ColumnPath;, access 
public
 Removed Method get, sig 
(Lparquet/hadoop/metadata/ColumnPath;Lparquet/schema/PrimitiveType$PrimitiveTypeName;Lparquet/hadoop/metadata/CompressionCodecName;Ljava/util/Set<Lparquet/column/Encoding;>;JJJJJ)Lparquet/hadoop/metadata/ColumnChunkMetaData;,
 desc 
(Lparquet/hadoop/metadata/ColumnPath;Lparquet/schema/PrimitiveType$PrimitiveTypeName;Lparquet/hadoop/metadata/CompressionCodecName;Ljava/util/Set;JJJJJ)Lparquet/hadoop/metadata/ColumnChunkMetaData;,
 access public static
Class parquet.hadoop.metadata.ColumnChunkProperties
 Added Method getPath, desc ()Lparquet/common/schema/ColumnPath;, access public
 Removed Method getPath, desc ()Lparquet/hadoop/metadata/ColumnPath;, access 
public
 Removed Method get, sig 
(Lparquet/hadoop/metadata/ColumnPath;Lparquet/schema/PrimitiveType$PrimitiveTypeName;Lparquet/hadoop/metadata/CompressionCodecName;Ljava/util/Set<Lparquet/column/Encoding;>;)Lparquet/hadoop/metadata/ColumnChunkProperties;,
 desc 
(Lparquet/hadoop/metadata/ColumnPath;Lparquet/schema/PrimitiveType$PrimitiveTypeName;Lparquet/hadoop/metadata/CompressionCodecName;Ljava/util/Set;)Lparquet/hadoop/metadata/ColumnChunkProperties;,
 access public static
 Added Method get, sig 
(Lparquet/common/schema/ColumnPath;Lparquet/schema/PrimitiveType$PrimitiveTypeName;Lparquet/hadoop/metadata/CompressionCodecName;Ljava/util/Set<Lparquet/column/Encoding;>;)Lparquet/hadoop/metadata/ColumnChunkProperties;,
 desc 
(Lparquet/common/schema/ColumnPath;Lparquet/schema/PrimitiveType$PrimitiveTypeName;Lparquet/hadoop/metadata/CompressionCodecName;Ljava/util/Set;)Lparquet/hadoop/metadata/ColumnChunkProperties;,
 access public static
Class parquet.hadoop.metadata.ColumnPath
 Removed Class , access final public super synchronized
{code}

> Release parquet-mr 1.6.0
> ------------------------
>
>                 Key: PARQUET-211
>                 URL: https://issues.apache.org/jira/browse/PARQUET-211
>             Project: Parquet
>          Issue Type: Bug
>          Components: parquet-mr
>    Affects Versions: 1.6.0
>            Reporter: Ryan Blue
>            Assignee: Ryan Blue
>             Fix For: 1.6.0
>
>
> Need to determine a list of tasks that should be done before release. Please 
> add issues as sub-tasks.



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

Reply via email to