org.apache.paimon.format.parquet.reader

Class NestedColumnReader

java.lang.Object

org.apache.paimon.format.parquet.reader.NestedColumnReader

All Implemented Interfaces:

ColumnReader<WritableColumnVector>

public class NestedColumnReader
extends Object
implements ColumnReader<WritableColumnVector>

This ColumnReader mainly used to read `Group` type in parquet such as `Map`, `Array`, `Row`. The method about how to resolve nested struct mainly refer to : The striping and assembly algorithms from the Dremel paper.

Brief explanation of reading repetition and definition levels: Repetition level equal to 0 means that this is the beginning of a new row. Other value means that we should add data to the current row.

For example, if we have the following data: repetition levels: 0,1,1,0,0,1,[0] (last 0 is implicit, normally will be the end of the page) values: a,b,c,d,e,f will consist of the sets of: (a, b, c), (d), (e, f).

Definition levels contains 3 situations: level = maxDefLevel means value exist and is not null level = maxDefLevel - 1 means value is null level < maxDefLevel - 1 means value doesn't exist For non-nullable (REQUIRED) fields the (level = maxDefLevel - 1) condition means non-existing value as well.

Quick example (maxDefLevel is 2): Read 3 rows out of: repetition levels: 0,1,0,1,1,0,0,... definition levels: 2,1,0,2,1,2,... values: a,b,c,d,e,f,... Resulting buffer: a,n, ,d,n,f that result is (a,n),(d,n),(f) where n means null

Constructor Summary

Constructors

Constructor and Description
NestedColumnReader(boolean isUtcTimestamp, org.apache.parquet.column.page.PageReadStore pages, ParquetField field)

Method Summary

Modifier and Type	Method and Description
void	readToVector(int readNumber, WritableColumnVector vector)

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

NestedColumnReader

public NestedColumnReader(boolean isUtcTimestamp,
                          org.apache.parquet.column.page.PageReadStore pages,
                          ParquetField field)

Method Detail

readToVector

public void readToVector(int readNumber,
                         WritableColumnVector vector)
                  throws IOException

Specified by:

readToVector in interface ColumnReader<WritableColumnVector>

Parameters:

readNumber - number to read.

vector - vector to write.

Throws:

IOException