Flight Aware has a great visualization for show flight delay across the USA. They call this the Misery Map
It also has a timeline that you can play across the last three days
This is a very cool way to do a depth first traversal of JSON that I have test in the Oracle 19c version of the database So the cool part is col3 clob format json with wrapper path '$.col3[*]' that stores a JSON objects for col3 so that the outer loop does not care what the object contents are the inner loop then parses this JSON. This is a very compact and readable way of traversing all of your JSON begin for rec1 in ( select col1, col2, col3 from json_table( '{"col1": 1, "col2" : 2, col3 : {"col3foo1" :4, "col3foo2": 5}}' , '$[*]' columns col1 varchar2 ( 200 ) path '$.col1' , col2 varchar2 ( 200 ) path '$.col2' , col3 clob format json with wrapper path '$.col3[*]' ) ) loop dbms_output.put_line( 'col1=' ...
Your REST Client can consume JSON produced from a REST API. You can use Oracle SQL JSON functions to construct a JSON object and you can then use this to abstract out the source of the data. Your REST API client does not need to know that the data came from a set of tables. Here is an example of how to construct a JSON object. WITH t AS ( SELECT JSON_OBJECT ( 'measureDetails' VALUE ( SELECT JSON_ARRAYAGG(JSON_OBJECT('measureName' VALUE b.measure_name, 'classifications' VALUE JSON_ARRAYAGG(JSON_OBJECT('classification' ...
Comments