See also CQL in the Glossary and RAML Module Builder’s CQL, CQL2PgJSON, and Post Tenant (schema.json) documentation sections for further CQL information. The Wiki FOLIOtips/Searching has notes about search indexes, and example CQL syntax search queries.

See explanations and examples below for Field match operator and Word match operators.

Field match operator: ==

The CQL “exact match” operator (==) is used to match against a complete field.

Truncation is enabled using the * wildcard, either on the right end or on the left end.

If the == search is to be accelerated, then create a B-tree database index by adding an "index" or "uniqueIndex" entry for the field in schema.json as explained in RAML Module Builder Post Tenant API documentation.

Only right end truncation is supported by B-tree database indexes.

Field match examples 1

Consider the CQL query:

field == "abc xyz"

It does match only:

abc xyz

It does not match:

The abc xyz
xyz abc
abc xyz.
abc, xyz

This has the SQL equivalents (both are the same):
table.field = 'abc xyz'
table.field LIKE 'abc xyz'

Field match examples 2

Consider the CQL query:

field == "abc xyz*"

This is a left bounded match with right truncation.

It does match:

abc xyz
abc xyz.
abc xyzq
abc xyz qqq

It does not match:

The abc xyz
xyz abc
abc, xyz

This has the SQL equivalent:
table.field LIKE 'abc xyz%'

Field match examples 3

Consider the CQL query:

field == "*abc xyz*"

This is left and right truncation.

It does match:

abc xyz
abc xyz.
abc xyzq
abc xyz qqq
The abc xyz
The abc xyzq
The abc xyz qqq

It does not match:

xyz abc
abc, xyz

This has the SQL equivalent:
table.field LIKE '%abc xyz%'

Note that this is slow on large datasets because b-tree database indexes support only right truncation.

Word match operators: =, adj, all, any

The four word match operators ignore punctuation and whitespace, and they match against words.

Truncation of a word is possible using the * wildcard, but only on the right, not on the left.

It is implemented using PostgreSQL’s to_tsvector @@ to_tsquery full text search.

all matches if each word of the query string exists somewhere.

any matches if any word of the query string exists somewhere.

adj matches if all words of the query string exist consecutively in that order, there may be any whitespace and punctuation in between.

= is a synonym for adj.

If a word match search is to be accelerated, then create a full text database index by adding a "fullTextIndex" entry for the field in schema.json as explained in RAML Module Builder Post Tenant API documentation.

Only right end truncation is supported by word match operators and full text database indexes.

Word match examples 1

Consider each of the CQL queries:

field all "abc"
field any "abc"
field adj "abc"
field = "abc"

Each does match:

abc
The abc xyz
?abc!xyz

Each does not match:

abcd

These have the SQL equivalent:
to_tsvector('simple', table.jsonb->>'field') @@ to_tsquery('simple', 'abc')

Word match examples 2

Consider the CQL query:

field all "abc xyz"

It does match:

abc xyz
xyz abc
The abc xyz qqq
abc, xyz.

It does not match:

abc xyzq

This has the SQL equivalent:
to_tsvector('simple', table.jsonb->>'field') @@ to_tsquery('simple', 'abc & xyz')

Word match examples 3

Consider the CQL query:

field any "abc xyz"

It does match:

abc
xyz
abc xyzq
abc xyz
abc, xyz.
xyz abc
The abc xyz qqq

It does not match:

xyzq
qqq

This has the SQL equivalent:
to_tsvector('simple', table.jsonb->>'field') @@ to_tsquery('simple', 'abc | xyz')

Word match examples 4

Consider the CQL query:

field = "abc xyz"

It does match:

abc xyz
The abc xyz qqq
abc, xyz.

It does not match:

abc
xyz
xyz abc
abc xyzq

This has the SQL equivalent:
to_tsvector('simple', table.jsonb->>'field') @@ to_tsquery('simple', 'abc <-> xyz')

Word match examples 5

Consider the CQL query:

field = "abc*"

It does match:

abc
abcdef
The abcdef xyz
The!abcdef?xyz
xyz abc

It does not match:

xyzabc

This has the SQL equivalent:
to_tsvector('simple', table.jsonb->>'field') @@ to_tsquery('simple', 'abc:*')