Always return NULL on error.
Adding also the file length.
modified: storage/connect/json.cpp
modified: storage/connect/jsonudf.cpp
- JSONCOL::WriteColumn Add types SHORT and BIGINT as accepted
modified: storage/connect/tabjson.cpp
- TDBJSN: Make this type use a separate storage for Json parsing
and retrieve this memory between each rows. This is necessary
to be able to handle big tables. See MDEV-9228.
modified: storage/connect/tabjson.cpp
modified: storage/connect/tabjson.h
Always return NULL on error.
Adding also the file length.
modified: storage/connect/json.cpp
modified: storage/connect/jsonudf.cpp
- JSONCOL::WriteColumn Add types SHORT and BIGINT as accepted
modified: storage/connect/tabjson.cpp
- TDBJSN: Make this type use a separate storage for Json parsing
and retrieve this memory between each rows. This is necessary
to be able to handle big tables. See MDEV-9228.
modified: storage/connect/tabjson.cpp
modified: storage/connect/tabjson.h
modified: storage/connect/filamzip.cpp
Avoid calling the wrong AddValue (Windows compiler error ???)
modified: storage/connect/json.h
Fix looping bug in JARRAY::AddValue for arrays having one value.
Fix potential crash in JVALUE::SetValue
modified: storage/connect/json.cpp
Many changes to fix bugs, enhance memory handling and modify Jpath.
In JSNX some functions have been re-written or added to handle new Jpath.
BMX was re-defined to avoid a different size between Windows an Linux.
Jbin memory was fixed to use the proper memory when adding values.
Default pretty value is now 0 instead of 2.
Filename is stored in BSON when IsJson == 2.
BSON size is added to memlen in CalcLen when applicable.
The order or parameter was switch in Json_Object_Grp.
File name argument must be a constant (temporary?)
Json_Set_Item now returns file name when applicable.
modified: storage/connect/jsonudf.cpp
modified: storage/connect/jsonudf.h
Include "mycat.h"
modified: storage/connect/mycat.cc
Udf_json test revisited and fixed for Linux
modified: storage/connect/mysql-test/connect/r/json_udf.result
modified: storage/connect/mysql-test/connect/t/json_udf.inc
modified: storage/connect/mysql-test/connect/t/json_udf.test
modified: storage/connect/mysql-test/connect/t/json_udf2.inc
was fixed in HA_CONNECT::CondFilter moving res= pval->val_str(&tmp)
but this was wrong. Now res is only used for strings.
Change version number
modified: storage/connect/ha_connect.cc
Add some new UDF's
modified: storage/connect/json.cpp
modified: storage/connect/json.h
modified: storage/connect/jsonudf.cpp
modified: storage/connect/jsonudf.h
Fix change in tests json_udf
modified: storage/connect/mysql-test/connect/r/json_udf.result
modified: storage/connect/mysql-test/connect/t/json_udf.inc
modified: storage/connect/mysql-test/connect/t/json_udf.test
modified: storage/connect/json.cpp
modified: storage/connect/json.h
modified: storage/connect/jsonudf.cpp
modified: storage/connect/mysql-test/connect/r/json_udf.result
modified: storage/connect/mysql-test/connect/t/json.test
modified: storage/connect/tabjson.cpp
Fix wrong calculation of Estimated Length when the table has virtual or special columns
modified: storage/connect/reldef.h
modified: storage/connect/tabdos.cpp
Fix wrong handling of null values in ODBCCOL::ReadColumn
modified: storage/connect/tabodbc.cpp
Fix crash when SetValue_char is called with a negative length value.
This can happen in odbconn.cpp when SQLFetch returns SQL_NO_TOTAL (-4) as length.
modified: storage/connect/odbconn.cpp
modified: storage/connect/value.cpp
From the previous branch:
commit eda4928ff122a0845baf5ade83b4aa29244a3a89
Author: Olivier Bertrand <bertrandop@gmail.com>
Date: Mon Mar 9 22:34:56 2015 +0100
- Add discovery to JSON tables
When columns are not defined, CONNECT analyses the json file to find column definitions.
This wors only on table that are an array of objects. Pair keys are used to generate the
column names and pair values are used for its definition. When the LEVEL option is defined
as a not null integer, the eventual JPATH is scanned up to the LEVEL value.
From the current one:
- Fix MDEV-7521 when column names are utf8 encoded (not a general multi-charset fix)
- Adds more to JSON discovery processing and UDF's
- Use PlugDup everywhere it replaces PlugSubAlloc + strcpy.
modified:
storage/connect/json.cpp
storage/connect/json.h
storage/connect/mysql-test/connect/r/json.result
storage/connect/mysql-test/connect/t/json.test
storage/connect/tabjson.cpp
storage/connect/tabjson.h
- Avoid crash when a partition table name pattern is ill formed
(such as using place holder %i instead of %s)
modified:
storage/connect/ha_connect.cc
