There is Javascript in my SQL
0 likes382 views
The document discusses the use of JSON in PostgreSQL, highlighting its role in transferring and storing semi-structured data within relational databases. It compares the json and jsonb data types, detailing their advantages and disadvantages, such as processing speed and data preservation. Additionally, it outlines how to create indices for JSON data and describes the integration of JavaScript functions in PostgreSQL through the plv8 extension.
![PL/v8
● Plv8.execute(sql,[args]);
● Plv8.prepare(sql,[args]);
– var stmt = plv8.prepare('Select * from tbl where col1=$1',['text']);
– var rows = stmt.execute('foss'); (returns a cursor);
– stmt.free(); _> frees the prepared statement
● Cursors operations
– fetch([nrows])
– move([nrows])
– close()](https://image.slidesharecdn.com/pgday-asia-2016-mark-there-is-javascript-in-my-sql-160326160641/85/There-is-Javascript-in-my-SQL-21-320.jpg)
![Plv8
● Plv8.return_next()
→ stores records in internal tuple store,
→ content of store return at end of function
● Can also return an array of JSON matching
expected rowtype,
● Plv8.elog([level],message[,msg2,...]);](https://image.slidesharecdn.com/pgday-asia-2016-mark-there-is-javascript-in-my-sql-160326160641/85/There-is-Javascript-in-my-SQL-22-320.jpg)
There is Javascript in my SQL
- 1. Postgres – There is JavaScript in my SQL!
- 2. About Me ● Solutions integrator at Jumping Bean – Developer & Trainer – Social Media ● Twitter @mxc4 ● Twitter @jumpingbeansa ● LinkedIn
- 3. What is JSON? ● JSON -> JavaScript Object Notation, ● Used to transfer data in JavaScript, ● Light weight way to marshal JavaScript objects, ● Widely used in REST services, ● Used as generic data transfer object for unstructured data
- 4. Why store JSON? ● Need to store semi-structured data ● Reduce time to market – Ability to define data requirements “in-flight ● Need to store data from REST API, ● All the arguments for NoSQL
- 5. Why Use Postgres for NoSQL? ● Postgres is fast ● Postgres is ACID compliant, ● Best of both worlds – benefits of relational database with those of NoSQL databases, ● Already invested in Postgres – Why learn something new?
- 6. JSON/JSONB Data Type ● Core type ● Two data types – JSON – 9.3 ● Stores exact copy of input text but must be valid JSON – JSONB – 9.4 ● Stores a binary representation of the original input ● Encoding issues: – JSON spec uses UTF-8 encoding, – Database may use different encoding may be incompatibilities
- 7. JSON/JSONB ● Pros – JSON ● Ensures its valid JSON, ● Copy of original, preserves white space as well as order of keys and duplicate keys – JSONB ● Faster to process, ● Supports indexing, ● Tokenised BTree ● ● Cons – JSON ● Must be parsed when used in processing functions, – JSONB ● Duplicate keys and unnecessary white space discarded
- 8. Three Types of Functions ● Operators, – Functions to extract elements from JSON ● Creation functions, – Create JSON from database rows, arrays ● Processing functions – Typically to convert from JSON to records or recordset
- 12. Demo Querying Twitter Users
- 18. Creating Index ● Can create GIN (Generalised inverted index) index on JSONB, ● Allows use of @ contains operator , ● E.G – Create index idx01 on docs using GIN(body); ● Can create index on specific fields with – create index inx02 on docs ((doc >'col1')); – Text type index→ ● Text based index, ● Select tuser 'location' from followers where tuser >'location' = “South Africa”→ → – create index inx03 on docs((doc 'col1')); – JSON type index→ ● Index on JSON data type ● select tuser->'location' from followers where tuser->'location' = '"South Africa"'::jsonb
- 20. JavaScript Postgres Functions ● Postgres functions can be written in JavaScript, – Install PLv8 extensions ● “apt-get install postgresql-9.4-plv8” – Enable extension with: ● “Create extensions plv8”
- 21. PL/v8 ● Plv8.execute(sql,[args]); ● Plv8.prepare(sql,[args]); – var stmt = plv8.prepare('Select * from tbl where col1=$1',['text']); – var rows = stmt.execute('foss'); (returns a cursor); – stmt.free(); _> frees the prepared statement ● Cursors operations – fetch([nrows]) – move([nrows]) – close()
- 22. Plv8 ● Plv8.return_next() → stores records in internal tuple store, → content of store return at end of function ● Can also return an array of JSON matching expected rowtype, ● Plv8.elog([level],message[,msg2,...]);
- 24. The End