ParadeDB Docs

Term queries look for exact token matches. A term query is like an exact string match, but at the token level. Unlike match or phrase queries, term queries treat the query string as a finalized token. This means that the query string is taken as-is, without any further tokenization or filtering. Term queries use the === operator. To understand exactly how it works, let’s consider the following two term queries:

-- Term query 1
SELECT description, rating, category
FROM mock_items
WHERE description === 'running';

-- Term query 2
SELECT description, rating, category
FROM mock_items
WHERE description === 'RUNNING';

import { search } from "@paradedb/drizzle-paradedb";

// Term query 1
await db
  .select({
    description: mockItems.description,
    rating: mockItems.rating,
    category: mockItems.category,
  })
  .from(mockItems)
  .where(search.term(mockItems.description, "running"));

// Term query 2
await db
  .select({
    description: mockItems.description,
    rating: mockItems.rating,
    category: mockItems.category,
  })
  .from(mockItems)
  .where(search.term(mockItems.description, "RUNNING"));

from paradedb import ParadeDB, Term

# Term query 1
MockItem.objects.filter(
    description=ParadeDB(Term('running'))
).values('description', 'rating', 'category')

# Term query 2
MockItem.objects.filter(
    description=ParadeDB(Term('RUNNING'))
).values('description', 'rating', 'category')

from sqlalchemy import select
from sqlalchemy.orm import Session
from paradedb.sqlalchemy import search

term_query_1 = (
    select(MockItem.description, MockItem.rating, MockItem.category)
    .where(search.term(MockItem.description, "running"))
)

term_query_2 = (
    select(MockItem.description, MockItem.rating, MockItem.category)
    .where(search.term(MockItem.description, "RUNNING"))
)

with Session(engine) as session:
    {
        "rows_query_1": session.execute(term_query_1).all(),
        "rows_query_2": session.execute(term_query_2).all(),
    }

# Term query 1
MockItem.search(:description)
        .term("running")
        .select(:description, :rating, :category)

# Term query 2
MockItem.search(:description)
        .term("RUNNING")
        .select(:description, :rating, :category)

await dbContext
    .MockItems.Where(item => EF.Functions.Term(item.Description, "running"))
    .Select(item => new { item.Description, item.Rating, item.Category })
    .ToListAsync();

await dbContext
    .MockItems.Where(item => EF.Functions.Term(item.Description, "RUNNING"))
    .Select(item => new { item.Description, item.Rating, item.Category })
    .ToListAsync();

The first query returns:

     description     | rating | category
---------------------+--------+----------
 Sleek running shoes |      5 | Footwear
(1 row)

However, the second query returns no results. This is because term queries look for exact matches, which includes case sensitivity, and there are no documents in the example dataset containing the token RUNNING.

All tokenizers besides the literal tokenizer lowercase tokens by default. Make sure to account for this when searching for a term.

If you are using === to do an exact string match on the original text, make sure that the text uses the literal tokenizer.

How It Works

Under the hood, === simply finds all documents where any of their tokens are an exact string match against the query token. A document’s tokens are determined by the field’s tokenizer and token filters, configured at index creation time.

Examples

Let’s consider a few more hypothetical documents to see whether they would be returned by the term query. These examples assume that index uses the default tokenizer and token filters, and that the term query is running.

Original Text	Tokens	Match	Reason	Related
Sleek running shoes	`sleek` `running` `shoes`	✅	Contains the token `running`.
Running shoes sleek	`sleek` `running` `shoes`	✅	Contains the token `running`.
SLeeK RUNNING ShOeS	`sleek` `running` `shoes`	✅	Contains the token `running`.	Lowercasing
Sleek run shoe	`sleek` `run` `shoe`	❌	Does not contain the token `running`.	Stemming
Sleke ruining shoez	`sleke` `ruining` `shoez`	❌	Does not contain the token `running`.	Fuzzy
White jogging shoes	`white` `jogging` `shoes`	❌	Does not contain the token `running`.

Term Set

Passing a text array to the right-hand side of === means “find all documents containing any one of these tokens.”

SELECT description, rating, category
FROM mock_items
WHERE description === ARRAY['shoes', 'running'];

import { search } from "@paradedb/drizzle-paradedb";

await db
  .select({
    description: mockItems.description,
    rating: mockItems.rating,
    category: mockItems.category,
  })
  .from(mockItems)
  .where(search.term(mockItems.description, ["shoes", "running"]));

from paradedb import ParadeDB, TermSet

MockItem.objects.filter(
    description=ParadeDB(TermSet('shoes', 'running'))
).values('description', 'rating', 'category')

from sqlalchemy import or_, select
from sqlalchemy.orm import Session
from paradedb.sqlalchemy import search

stmt = (
    select(MockItem.description, MockItem.rating, MockItem.category)
    .where(
        or_(
            search.term(MockItem.description, "shoes"),
            search.term(MockItem.description, "running"),
        )
    )
)

with Session(engine) as session:
    session.execute(stmt).all()

MockItem.search(:description)
        .term_set("shoes", "running")
        .select(:description, :rating, :category)

await dbContext
    .MockItems.Where(item => EF.Functions.Term(item.Description, new[] { "shoes", "running" }))
    .Select(item => new { item.Description, item.Rating, item.Category })
    .ToListAsync();

​How It Works

​Examples

​Term Set

How It Works

Examples

Term Set