SentinelDB supports a subset of the SQL syntax. Below are a few examples of how to use SQL queries. They are sent to a special SQL API endpoint, or can be used in the dashboard.

Select statements


SELECT *, foo, bar
FROM "3aa42837-6de4-4203-a8c9-b9696d4408eb".records.None
WHERE foo=aaa AND bar=bbb
ORDER BY foo DESC, bar


  • SELECT clause can contain * and any other literal.

If object is record special select words are ‘‘ , ‘id’, ‘ownerid’, ‘updated’
If object is user special select words are ‘
‘, ‘id’, ‘updated’, ‘username’, ‘password’, ’email’, ‘status’
‘*’ extracts the whole user/record represented as json.
these words extract fields of the object itself and not its body that comes from the client.
All other select words extract info from the body of the object. As it is json dot notation can be used. Example: foo.bar.bass

  • FROM clause must be in the form .users or .records. Default type name is None (when no type is provided None is used)
  • WHERE clause supports list of conditions separated by AND word. Each condition is in the form field=value.
    As in SELECT clause special words conditions are checked against the object fields and not against the body fields.
    Other fields conditions are checked against object’s body only if they are made searchable (with search schema)
  • ORDER BY clause can contain list of fields separated by ‘,’ ASC and DESC sorting is allowed for every field
  • LIMIT clause expects numeric value for limiting results

Update statements


UPDATE "3aa42837-6de4-4203-a8c9-b9696d4408eb".records.None
   SET 'some.thing'=something
   WHERE id='953df9c1-5917-4ff0-a5e0-604c07071324'


  • UPDATE clause is the same as FROM clause from Select statemet
  • SET clause contains list of separated by ‘,’
    It uses same special words as select clause. If fields are not special dot notation creates new json structure in the body:
    SET 'foo.bar'=green creates {"foo":{"bar" : "green"}}
  • WHERE clause is the same like in Select statment. If it contains only id it is fast, otherwise all matching objects are searched so it’s slower.

Delete statements


DELETE "3aa42837-6de4-4203-a8c9-b9696d4408eb".records.None
    WHERE id='953df9c1-5917-4ff0-a5e0-604c07071324'


  • DELETE clause is the same like UPDATE and FROM
  • WHERE is the same like in UPDATE statement including fast processing by id

Insert statements


INSERT INTO "3aa42837-6de4-4203-a8c9-b9696d4408eb".records.None
    (ownerId, foo, bar)
    VALUES ('4aa42837-6de4-4203-a8c9-b9696d4408eb', green, yellow)


  • INSERT clause is the same like UPDATE and FROM + added list of field names. Special field names are the same like in other clauses
  • VALUES clasue contains the values of the fields in the INSERT clause

Common syntax

When field or value contains ‘-‘, ‘.’ or other special characters it must be quoted with ” or ‘ In other cases quotes are not mandatory, but can be used.

Prepared statements

All SQL API endpoints (select, update, delete, insert) support prepared statements.
To keep APIs stateless parameter values are always expected with the query itself. Number of ‘?’ and number of params must be the same



{"query":"select  id,foo,bar, updated, ownerId from '3aa42837-6de4-4203-a8c9-b9696d4408eb'.records.None where id=?  and updated > ?",
 "params":["953df9c1-5917-4ff0-a5e0-604c07071324", ""]


{"query":"update '3aa42837-6de4-4203-a8c9-b9696d4408eb'.records.None set 'some.thing'=? where id=?",
 "params" :["foo","5095dfb5-5bbf-4a15-8d7d-12d66a3a4f2c"]


{"query":"delete '3aa42837-6de4-4203-a8c9-b9696d4408eb'.records.None where id=?",


{"query":"insert into '3aa42837-6de4-4203-a8c9-b9696d4408eb'.records.None (foo, bar, 'bass.bar') values (?,?,?) ",
 "params":["pink", "green", "yellow"]