Query Parser - ParadeDB

The parse query accepts a Tantivy query string. The intended use case is for accepting raw query strings provided by the end user. To use it, pass the table name or alias to the left-hand side of @@@ and pdb.parse('<query>') to the right-hand side. Passing an indexed column with pdb.parse() also remains supported.

SELECT description, rating, category FROM mock_items
WHERE mock_items @@@ pdb.parse('description:(sleek shoes) AND rating:>3');

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

await db
  .select({
    description: mockItems.description,
    rating: mockItems.rating,
    category: mockItems.category,
  })
  .from(mockItems)
  .where(search.parse(mockItems.id, "description:(sleek shoes) AND rating:>3"));

from paradedb import ParadeDB, Parse

MockItem.objects.filter(
    id=ParadeDB(Parse('description:(sleek shoes) AND rating:>3'))
).values('description', 'rating', 'category')

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

stmt = (
    select(MockItem.description, MockItem.rating, MockItem.category)
    .where(search.parse(MockItem.id, "description:(sleek shoes) AND rating:>3"))
)

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

MockItem.search(:id)
        .parse("description:(sleek shoes) AND rating:>3")
        .select(:description, :rating, :category)

await dbContext
    .MockItems.Where(item =>
        EF.Functions.Parse(item.Id, "description:(sleek shoes) AND rating:>3")
    )
    .Select(item => new { item.Description, item.Rating, item.Category })
    .ToListAsync();

Please refer to the Tantivy docs for an overview of the query string language.

Lenient Parsing

By default, strict syntax parsing is used. If the query has invalid syntax or names an unknown field, it fails. By setting lenient to true, the query is executed on a best-effort basis, ignoring invalid parts. Terms without a field name search across indexed fields in both modes.

SELECT description, rating, category FROM mock_items
WHERE mock_items @@@ pdb.parse('description:(sleek shoes) AND rating:>3', lenient => true);

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

await db
  .select({
    description: mockItems.description,
    rating: mockItems.rating,
    category: mockItems.category,
  })
  .from(mockItems)
  .where(
    search.parse(mockItems.id, "description:(sleek shoes) AND rating:>3", {
      lenient: true,
    }),
  );

from paradedb import ParadeDB, Parse

MockItem.objects.filter(
    id=ParadeDB(Parse('description:(sleek shoes) AND rating:>3', lenient=True))
).values('description', 'rating', 'category')

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

stmt = (
    select(MockItem.description, MockItem.rating, MockItem.category)
    .where(search.parse(MockItem.id, "description:(sleek shoes) AND rating:>3", lenient=True))
)

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

MockItem.search(:id)
        .parse("description:(sleek shoes) AND rating:>3", lenient: true)
        .select(:description, :rating, :category)

await dbContext
    .MockItems.Where(item =>
        EF.Functions.Parse(
            item.Id,
            "description:(sleek shoes) AND rating:>3",
            lenient: true
        )
    )
    .Select(item => new { item.Description, item.Rating, item.Category })
    .ToListAsync();

Conjunction Mode

By default, terms in the query string are ORed together. With conjunction_mode set to true, they are instead ANDed together. For instance, the following query returns documents containing both sleek and shoes.

SELECT description, rating, category FROM mock_items
WHERE mock_items @@@ pdb.parse('description:(sleek shoes)', conjunction_mode => true);

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

await db
  .select({
    description: mockItems.description,
    rating: mockItems.rating,
    category: mockItems.category,
  })
  .from(mockItems)
  .where(
    search.parse(mockItems.id, "description:(sleek shoes)", {
      conjunctionMode: true,
    }),
  );

from paradedb import ParadeDB, Parse

MockItem.objects.filter(
    id=ParadeDB(Parse('description:(sleek shoes)', conjunction_mode=True))
).values('description', 'rating', 'category')

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

stmt = (
    select(MockItem.description, MockItem.rating, MockItem.category)
    .where(search.parse(MockItem.id, "description:(sleek shoes)", conjunction_mode=True))
)

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

MockItem.search(:id)
        .parse("description:(sleek shoes)", conjunction_mode: true)
        .select(:description, :rating, :category)

await dbContext
    .MockItems.Where(item =>
        EF.Functions.Parse(
            item.Id,
            "description:(sleek shoes)",
            null,
            true
        )
    )
    .Select(item => new { item.Description, item.Rating, item.Category })
    .ToListAsync();