MySQL JSON Functions

Copyright © 2013, Oracle and/or its affiliates. All rights reserved.1


Insert Picture HereMySQL JSON FunctionsSveta SmirnovaPrincipal Technical Support Engineer


Program Agenda

Introduction: NoSQL involvement on MySQL Overview of the functions Function descriptions Where to get The future


Insert Chart HereMain historic points of NoSQL

1997 1998 2000 2003 2004 2008 2009 2010 2012 2013

0

0.5

1

1.5

2

2.5

3

3.5

4

Databases

NoSQL history

NoSQLdatabase

NoSQLterm

db40 memcachedBigTable

MemcacheDBTarantoolCouchDB

MongoDBRedis

WebSphere

VirtuosoDynamoDB

Hbase


Insert Chart HereMain points for NoSQL features in MySQL

20092010

20112012

2013

0

1

2

3

4

5

6

NoSQL in MySQL world

Memcached bugsat mysql.com

HandlerSocket

InnoDB with memcachedMemcache API for MySQL Cluster

NoSQL Connectorfor JavaScriptEXPLAIN in JSON

Hadoop Appliermysqlv8udfsJSON UDFs


Insert Chart HereA little bit before this year

2012

2013

00.020.040.060.080.1

0.120.140.160.180.2

JSON functions in MySQL

Version 0.1Never published

Version 0.2


JSON functions overview

FunctionsManipulate JSON text

● Validate● Search● Modify

UDF functions● Easy to install● Independent from MySQL server version

Work on all MySQL supported platformsBinaries for Linux, Mac OSX 7 and Windows

What are they doing?


How to install?UNIX

create function json_valid returns integersoname 'libmy_json_udf.so';create function json_search returns stringsoname 'libmy_json_udf.so';create function json_extract returns stringsoname 'libmy_json_udf.so';create function json_replace returns stringsoname 'libmy_json_udf.so';create function json_append returns stringsoname 'libmy_json_udf.so';...


How to install?Windows

create function json_remove returns string soname 'my_json_udf.dll';create function json_set returns string soname 'my_json_udf.dll';create function json_merge returns string soname 'my_json_udf.dll';create function json_contains_key returns integersoname 'my_json_udf.dll';create function json_test_parser returns stringsoname 'my_json_udf.dll';...


Dependencies

Regex libraryUNIX

● Usually already existsWindows

● Provided with sources● Compiled statically

You don't need to install additional libraries to run the UDF!


How to compile?UNIX

You need:● cmake● Regex library (usually already exists)● Working compiler

To build:● cmake . DMYSQL_DIR=/home/sveta/src/mysql5.5● make


How to compile?Windows

You need: PCRE static libraries, Visual Studio and cmake (cmake.org)To build PCRE:

● Download sources from http://www.pcre.org/ Recommended version is 8.33 or cd pcre8.33● Unpack archive and run:

● "C:\Program Files (x86)\CMake 2.8\bin\cmake.exe" G "Visual Studio 11 Win64"

● devenv PCRE.sln /build Release

● To build JSON UDFs copy pcre.h, pcreposix.h, Release/pcre.lib, Release/pcreposix.lib into JSON UDFs source directory, then run:

● "C:\Program Files (x86)\CMake 2.8\bin\cmake.exe" G "Visual Studio 11 Win64" . DMYSQL_DIR="C:/MySQL/mysql5.5"

● devenv my_json_udf.sln /build Release

http://www.pcre.org/


Disadvantages of UDFs

They are slow● Can not use certain server features, available for internal functions

● Full-text

● Indexes

● Built-in optimization


Why UDFs?

Flexibility● You can install as many functions as you want: single, few or all of them

Compatible with any server version● You don't need to upgrade to not stable version only to try these functions

Easy to addEasy to changeEasy to removeFeature requests are easy to implement

● Report bugs!● Raise your opinion!


Insert Picture Here

Functions descriptions


json_valid(doc)

Checks if doc is valid JSON document.Returns 1 if document is valid, 0 if document is invalid.Strict format as described at http://json.org

http://json.org/


json_valid(doc)

mysql> select json_valid('{"MySQL connect": ["conference", 2013]}');++| json_valid('{"MySQL connect": ["conference", 2013]}') |++| 1 |++1 row in set (0.01 sec)


json_valid(doc)

mysql> select json_valid('{"MySQL connect"}');++| json_valid('{"MySQL connect"}') |++| 0 |++1 row in set (0.00 sec)


json_contains_key(doc, keypart1, keypart2, ...)

Checks if documents contains specified key.Returns 1 if key exists, 0 if not exists or NULL if parsing failed.Warning! This version does not check whole document for validity.



SET optimizer_trace=1;mysql> select user from mysql.user;++...mysql> select json_contains_key(trace, 'steps', '0', 'join_optimization', 'steps', '0', 'condition_processing') as contains from information_schema.optimizer_trace;++| contains |++| 0 |++1 row in set (0.01 sec)



mysql> select user from mysql.user where user='Sveta';++...mysql> select json_contains_key(trace, 'steps', '0', 'join_optimization', 'steps', '0', 'condition_processing') as contains from information_schema.optimizer_trace;++| contains |++| 1 |++1 row in set (0.01 sec)


json_extract(doc, keypart1, keypart2, ...)

Extracts value of the specified key.Returns value of the key specified, NULL if the key does not exist or if parsing failed.Warning! This version does not check whole document for validity.



SET optimizer_trace=1;mysql> select user from mysql.user;++...mysql> select json_extract(trace, 'steps', '0', 'join_optimization', 'steps', '0', 'condition_processing') as value from information_schema.optimizer_trace;++| value |++| NULL |++1 row in set (0.01 sec)


Search path

{“steps”: [ {“join_optimization”: {“steps”: [ {“condition_processing”: .....

json_extract(trace, 'steps', '0', 'join_optimization', 'steps', '0', 'condition_processing')



mysql> select user from mysql.user where user='Sveta';++...mysql> select json_extract(trace, 'steps', '0', 'join_optimization', 'steps', '0', 'condition_processing') as value from information_schema.optimizer_trace\G*************************** 1. row ***************************value: { "condition": "WHERE", "original_condition": "(`mysql`.`user`.`User` = 'sveta')", "steps": [....


json_append(doc, keypart1, keypart2, ..., new_element)

Inserts new element into JSON document.Returns document with appended element, original document if no place to insert or NULL if parsing failed.Warning! This version does not check whole document for validity.



mysql> select json_append('{"MySQL connect": ["conference", 2013]}', 'MySQL connect', '2', '”San Francisco”') as 'MySQL connect';++| MySQL connect |++| {"MySQL connect": ["conference", 2013, “San Francisco”]} |++1 row in set (0.00 sec)



mysql> select json_append('{"MySQL connect": ["conference", 2013]}', 'MySQL connect', '1', '”San Francisco”') as 'MySQL connect';++| MySQL connect |++| {"MySQL connect": ["conference", 2013]} |++1 row in set (0.00 sec)


json_replace(doc, keypart1, keypart2, ..., new_value)

Updates value of the specified key.Returns document with replaced key or original document if no such an element found, NULL if parsing failed.Warning! This version does not check whole document for validity.



mysql> select json_replace('{"MySQL connect": ["conference", 2013]}', 'MySQL connect', '0', '"User conference"') as 'MySQL connect';++| MySQL connect |++| {"MySQL connect": ["User conference", 2013]} |++1 row in set (0.00 sec)



mysql> select json_replace('{"MySQL connect": ["conference", 2013]}', 'MySQL connect', '2', '"User conference"') as 'MySQL connect';++| MySQL connect |++| {"MySQL connect": ["conference", 2013]} |++1 row in set (0.00 sec)


json_set(doc, keypart1, keypart2, ..., new_value)

Performs kind of INSERT ... ON DUPLICATE KEY UPDATE operation.Returns document with updated or inserted element or NULL if parsing failed. Warning! This version does not check whole document for validity.



mysql> select json_set('{"MySQL connect": ["conference", 2013]}', 'MySQL connect', '0', '"User conference"') as 'MySQL connect';++| MySQL connect |++| {"MySQL connect": ["User conference", 2013]} |++1 row in set (0.00 sec)



mysql> select json_set('{"MySQL connect": ["conference", 2013]}', 'MySQL connect', '2', '"San Francisco"') as 'MySQL connect';++| MySQL connect |++| {"MySQL connect": ["conference", 2013, "San Francisco"]} |++1 row in set (0.00 sec)


json_remove(doc, keypart1, keypart2, ...)

Removes element specified by the key.Returns document without the element, original document if no element found or NULL if parsing failed. Warning! This version does not check whole document for validity.



mysql> select json_remove('{"MySQL connect": ["conference", 2013]}', 'MySQL connect', '1') as 'MySQL connect';++| MySQL connect |++| {"MySQL connect": ["conference"]} |++1 row in set (0.00 sec)



mysql> select json_remove('{"MySQL connect": ["conference", 2013]}', 'MySQL connect', '2') as 'MySQL connect';++| MySQL connect |++| {"MySQL connect": ["conference", 2013]} |++1 row in set (0.00 sec)


json_search(doc, value)

Searches for specified value in the document.Returns key path of the element which contains the value in reverse order or NULL if not found or parsing failed.Warning! This version does not check whole document for validity.



mysql> select json_search(trace, '"trivial_condition_removal"') from information_schema.optimizer_trace;+| json_search(trace, '"trivial_condition_removal"') |++| transformation:0:steps:condition_processing:0:steps:join_optimization:0:steps:: |++1 row in set (0.00 sec)



mysql> select json_search(trace, '"trivial_condition"') from information_schema.optimizer_trace;++| json_search(trace, '"trivial_condition"') |++| NULL |++1 row in set (0.01 sec)


json_merge(doc1, doc2, ...)

Merges 2 or more documents into one.Returns first document with following documents appended.Warning! This version does not check whole document for validity.If one of following documents does not contain an opening curly bracket first documents merged are returned and warning is generated.NULL if first document does not contain an opening curly bracket.



mysql> select json_merge('{"MySQL connect": ["conference", 2012]}', '{"MySQL connect": ["conference", 2013]}') as 'MySQL connect';++| MySQL connect |++| {"MySQL connect": ["conference", 2012], "MySQL connect": ["conference", 2013]} |++1 row in set (0.00 sec)



mysql> select json_merge('{"MySQL connect": ["conference", 2012]}', '{"MySQL connect": ["conference", 2013]}', '1') as 'MySQL connect';++| MySQL connect |++| {"MySQL connect": ["conference", 2012], "MySQL connect": ["conference", 2013]} |++1 row in set (0.00 sec)


json_test_parser(doc)

Returns text representation of parse tree of the JSON document, partial parse tree or empty string if document is invalid.This function is supposed to use for tests only and should not be used in production.



mysql> select json_test_parser('{"MySQL connect": ["conference", 2013]}') as 'Parse tree'\G********************** 1. row **********************Parse tree: => "conference"; => 2013; "MySQL connect" => ["conference", 2013]; => {"MySQL connect": ["conference", 2013]};



mysql> select json_test_parser('{"MySQL connect": ["conference", 2013]') as 'Parse tree';++| Parse tree |++| => "conference"; => 2013;"MySQL connect" => ["conference", 2013]; |++1 row in set (0.00 sec)


Insert Picture Here

Where to get


Source code and binaries

MySQL Labs● Source code● Binaries

● x86 and x86_64

● Generic Linux

● Mac OSX 10.7

● Windows 7

● http://labs.mysql.com/

http://labs.mysql.com/


More information

Manuals and articles● README file● My blog: https://blogs.oracle.com/svetasmirnova/

Announces● My twitter: https://twitter.com/#!/svetsmirnova

https://blogs.oracle.com/svetasmirnova/

https://twitter.com/#!/svetsmirnova


Insert Picture Here

The future


Depends from you

Send bugsSend feature requestsMore you send – more ideas we implementNow you can affect decisions


Where to report bugs and feature requests

MySQL Community bugs database● https://bugs.mysql.com/● No special category for now● General category

● “MySQL Server: User-defined functions (UDF)”

Internal Oracle bugs database

● Ask MySQL Support engineer to open a bug report for you

● Category “UDFJSON”

https://bugs.mysql.com/


More NoSQLSessions at MySQL Connect

MySQL and Hadoop: Big Data Integration—Unlocking New Insights [CON2053, Sunday, 11:30 AM]MySQL User-Defined Functions....in JavaScript! [CON1738, passed]Oracle NoSQL Database: When Is It the Right Tool for the Job? [CON13034, Saturday, 4:00 PM]MySQL As a NoSQL Store with the InnoDB/memcached Plug-in [CON3457, Sunday, 1:00 PM]MySQL’s EXPLAIN Command New Features [HOL9734, passed]Big Data with MySQL and Hadoop [CON2342, Sunday, 5:30 PM]


References

https://blogs.oracle.com/svetasmirnova/https://twitter.com/#!/svetsmirnovahttp://json.org/http://www.pcre.org/http://dev.mysql.com/doc/refman/5.6/en/adding-functions.htmlhttp://bugs.mysql.com/https://support.oracle.com

https://blogs.oracle.com/svetasmirnova/

https://twitter.com/#!/svetsmirnova

http://json.org/

http://www.pcre.org/

http://dev.mysql.com/doc/refman/5.6/en/adding-functions.html

http://bugs.mysql.com/

https://support.oracle.com/


Insert Picture Here

?


Insert Picture Here

THANK YOU!


Graphic Section Divider


The preceding is intended to outline our general product direction. It is intended for information purposes only, and may not be incorporated into any contract.It is not a commitment to deliver any material, code, or functionality, and should not be relied upon in making purchasing decisions. The development, release, and timing of any features or functionality described for Oracle’s products remains at the sole discretion of Oracle.

MySQL JSON Functions

Technology