thevar1able/ClickHouse
1
0
Fork 0
mirror of https://github.com/ClickHouse/ClickHouse.git synced 2024-11-22 23:52:03 +00:00
Code Issues Packages Projects Releases Wiki Activity

Added comments [#CLICKHOUSE-2838].

Browse Source
This commit is contained in:
Alexey Milovidov 2017-12-01 17:15:38 +03:00
parent d4089695ad
commit 6fb18a7dc2
2 changed files with 30 additions and 15 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

44
dbms/src/DataTypes/IDataType.h
View File

@ -66,23 +66,26 @@ public:
virtual DataTypePtr clone() const = 0;
/** Binary serialization for range of values in column - for writing to disk/network, etc.
* 'offset' and 'limit' are used to specify range.
* limit = 0 - means no limit.
* offset must be not greater than size of column.
* offset + limit could be greater than size of column
* - in that case, column is serialized to the end.
*/
virtual void serializeBinaryBulk(const IColumn & column, WriteBuffer & ostr, size_t offset, size_t limit) const;
/** Read no more than limit values and append them into column.
* avg_value_size_hint - if not zero, may be used to avoid reallocations while reading column of String type.
*/
virtual void deserializeBinaryBulk(IColumn & column, ReadBuffer & istr, size_t limit, double avg_value_size_hint) const;
/** Some data types are represented in multiple streams while being serialized.
*
* Some data types are represented in multiple streams while being serialized.
* Example:
* - Arrays are represented as stream of all elements and stream of array sizes.
* - Nullable types are represented as stream of values (with unspecified values in place of NULLs) and stream of NULL flags.
*
* Different streams are identified by "path".
* If the data type require single stream (it's true for most of data types), the stream will have empty path.
* Otherwise, the path can have components like "array elements", "array sizes", etc.
*
* For multidimensional arrays, path can have arbiraty length.
* As an example, for 2-dimensional arrays of numbers we have at least three streams:
* - array sizes; (sizes of top level arrays)
* - array elements / array sizes; (sizes of second level (nested) arrays)
* - array elements / array elements; (the most deep elements, placed contiguously)
*
* Descendants must override either serializeBinaryBulk, deserializeBinaryBulk methods (for simple cases with single stream)
* or serializeBinaryBulkWithMultipleStreams, deserializeBinaryBulkWithMultipleStreams, enumerateStreams methods (for cases with multiple streams).
*
* Default implementations of ...WithMultipleStreams methods will call serializeBinaryBulk, deserializeBinaryBulk for single stream.
*/
struct Substream
@ -111,6 +114,12 @@ public:
using OutputStreamGetter = std::function<WriteBuffer*(const SubstreamPath &)>;
using InputStreamGetter = std::function<ReadBuffer*(const SubstreamPath &)>;
/** 'offset' and 'limit' are used to specify range.
* limit = 0 - means no limit.
* offset must be not greater than size of column.
* offset + limit could be greater than size of column
* - in that case, column is serialized till the end.
*/
virtual void serializeBinaryBulkWithMultipleStreams(
const IColumn & column,
OutputStreamGetter getter,
@ -123,6 +132,9 @@ public:
serializeBinaryBulk(column, *stream, offset, limit);
}
/** Read no more than limit values and append them into column.
* avg_value_size_hint - if not zero, may be used to avoid reallocations while reading column of String type.
*/
virtual void deserializeBinaryBulkWithMultipleStreams(
IColumn & column,
InputStreamGetter getter,
@ -135,6 +147,10 @@ public:
deserializeBinaryBulk(column, *stream, limit, avg_value_size_hint);
}
/** Override these methods for data types that require just single stream (most of data types).
*/
virtual void serializeBinaryBulk(const IColumn & column, WriteBuffer & ostr, size_t offset, size_t limit) const;
virtual void deserializeBinaryBulk(IColumn & column, ReadBuffer & istr, size_t limit, double avg_value_size_hint) const;
/** Serialization/deserialization of individual values.
*

1
dbms/src/Storages/MergeTree/MergeTreePartChecker.cpp
View File

@ -74,7 +74,6 @@ public:
size_t read(size_t rows)
{
ColumnPtr column = type->createColumn();
// TODO type->deserializeBinaryBulk(*column, uncompressed_hashing_buf, rows, 0);
return column->size();
}