Sort_result Class Reference

#include <sql_sort.h>

Public Member Functions
	Sort_result ()
bool	has_result_in_memory () const
bool	has_result () const

Public Attributes
IO_CACHE *	io_cache {nullptr}
bool	sorted_result_in_fsbuf {false}
	If the entire result fits in memory, we skip the merge phase. More...
unique_ptr_my_free< uchar >	sorted_result {nullptr}
uchar *	sorted_result_end {nullptr}
ha_rows	found_records {0}
	How many records in sort. More...

Constructor & Destructor Documentation

Sort_result()

Sort_result::Sort_result ( )

inline

Member Function Documentation

has_result()

bool Sort_result::has_result ( ) const

inline

has_result_in_memory()

bool Sort_result::has_result_in_memory ( ) const

inline

Member Data Documentation

found_records

ha_rows Sort_result::found_records {0}

How many records in sort.

io_cache

IO_CACHE* Sort_result::io_cache {nullptr}

sorted_result

unique_ptr_my_free<uchar> Sort_result::sorted_result {nullptr}

sorted_result_end

uchar* Sort_result::sorted_result_end {nullptr}

sorted_result_in_fsbuf

bool Sort_result::sorted_result_in_fsbuf {false}

If the entire result fits in memory, we skip the merge phase.

We may leave the result in the parent Filesort_info's filesort_buffer (indicated by sorted_result_in_fsbuf), or we may strip away the sort keys, and copy the sorted result into a new buffer. Unique always uses the latter. This new buffer is [sorted_result ... sorted_result_end]

See also

save_index()

---

The documentation for this class was generated from the following file:

• sql/sql_sort.h