SQLite FTS4: A query containing a single unary not expression? - sql

The SQLite3 FTS4 documentation regarding the standard query unary - operator states:
An FTS query may not consist entirely of terms or term-prefix queries
with unary "-" operators attached to them.
Is there any way around this? Specifically, I want to retrieve all the rows in my table that do not contain a term or term-prefix.
sqlite> select * from search;
1|Clock
2|Un
sqlite> select id from search where text match 'Un';
2
sqlite> select id from search where text match 'clock -Un';
1
sqlite> select id from search where text match '-Un';
Error: malformed MATCH expression: [-Un]
sqlite> select id from search where text match '-Un*';
Error: malformed MATCH expression: [-Un*]

select id
from search
where docid not in (select docid
from search
where text match 'Un')

Related

SQL String Pattern Matching

I am trying to run a simple select query on a table to exclude all the records containing a single word on Snowflake.
For E.g. there is a column - " NAME" having datatype - STRING - containing a single word and combination of words.
I want to build a query to exclude all iterations of a sub-string present in that column ( uppercase,lowercase and camel-case via a single query)
I am trying to run a simple select query on a table to exclude all the records containing a single word
You could use ILIKE to perform case-insensitive comparison:
SELECT *
FROM tab
WHERE NAME NOT ILIKE '%phrase%';
db<>fiddle demo

sql query with multiple partial match condition

i have a table column looks like below.
what is the sql query statement i can use to have multiple partial match conditions?
search by ID or Name
if search abc then list the row A1 , row A2
if search test then list the row A1 , row A2, row 3
if search ghj then list the row A2
i was trying this but nothing return:
SELECT * FROM table where colB LIKE '"ID":"%abc%"'
updating data in text
{"ItemId":"123","IDs":[{"ID":"abc","CodingSystem":"cs1"}],"Name":"test itemgh"}
{"ItemId":"123","IDs":[{"ID":"ghj","CodingSystem":"cs1"}],"Name":"test abc"}
{"ItemId":"123","IDs":[{"ID":"defg","CodingSystem":"cs1"}],"Name":"test 111"}
JSON parsing
Oracle
Looked into the JSON parsing capabilities of Oracle and I managed to make running a query like this:
select * from table t where json_exists(t.colB, '$.IDs[?(#.ID=="abc")]') or json_exists(t.colB, '$.IDs?(#.name=="abc"')
And inside the same JSON query expression:
select * from table t where json_exists(t.colB, '$.IDs[?(#.ID=="abc" || #.name=="abc")]')
The call of function json_exists() is the key to this.
The first parameter can be a VARCHAR2, and I also tried with a BLOB containing text, and it works.
The second parameter is the path to your json object attribute that needs to be tested, with the condition.
I wrote two ORed conditions for the ID and for the Name, but maybe there is a better JSON query expression you can use to include them both.
More information about json_exists() function here.
Postgres
There is a JSON datatype in Postgres that supports parsing in queries.
So, if your colB column is declared as JSON you can do something like this:
select * from table where colB->>'Name' LIKE '%abc%';
And in order to have available the array elements of the IDs array, you should use the function json_array_elements().
select * from table, json_array_elements(colB->'IDs') e where colB->>'Name' LIKE '%abc%' or e->>'ID' = 'abc';
Check an example I created for you here.
Here is an online tool for online testing your JSON queries.
Check also this question in SO.
MSSQL Server 2017
I made a couple of tests also with MS SQL Server, and I managed to create an example searching for partial matching in the name field.
select * from table where JSON_VALUE(colB,'$.Name') LIKE '%abc%';
And finally I arrived to a working query that does partial match to the Name field and full match to the ID field like this:
select * from table t
CROSS APPLY OPENJSON(colB, '$.IDs') WITH (
ID VARCHAR(10),
CodingSystem VARCHAR(10)
) e
where JSON_VALUE(t.colB,'$.Name') LIKE '%abc%'
or e.ID = 'abc';
The problem is that we need to open the IDs array, and make something like a table from it, that can be queried also by accessing its columns.
The example I created is here.
LIKE text query
Your tries are good but you misplace the % symbols. They have to be first and last in your given string:
If you want the ID to be the given value:
SELECT * FROM table where colB LIKE '%"ID":"abc"%'
If the given value can be anywhere, then don't put the "ID" part:
SELECT * FROM table where colB LIKE '%abc%'
If the given value can be only on the ID or Name field then:
SELECT * FROM table where colB LIKE '%"ID":"abc"%' OR colB LIKE '%"Name":"abc"%'
And because you are giving hard-coded identifiers of fields (eg ID and Name) that can be in variable case:
SELECT * FROM table where lower(colB) LIKE '%"id":"abc"%' OR lower(colB) LIKE '%"name":"abc"%'
Assuming that the number of spaces do not vary between the : character and the value or the name of the properties.
For partial matching you can use more % in between like '%"name":"%abc%"%':
SELECT * FROM table where lower(colB) LIKE '%"id":"abc"%' OR lower(colB) LIKE '%"name":"%abc%"%'
Regular Expressions
A different option would be to test with regular expressions.
Consider checking this: Oracle extract json fields using regular expression with oracle regexp_substr

MariaDB MATCH AGAINST with single quote search term?

I'm trying to find a working query for using MATCH AGAINST while having a search term containing a single quote.
Example data in the database table:
I'm a freak
Example search term:
I'm
Search queries I tried:
SELECT * FROM table WHERE MATCH (name) AGAINST ('"I\'m"' IN NATURAL LANGUAGE MODE);
SELECT * FROM table WHERE MATCH (name) AGAINST ('"I\'m"' IN BOOLEAN MODE);
SELECT * FROM table WHERE MATCH (name) AGAINST ('I\'m*' IN BOOLEAN MODE);
SELECT * FROM table WHERE MATCH (name) AGAINST ('(I\'m)*' IN BOOLEAN MODE);
...and many more. Nothing is working.
I'm using MariaDB 10.1.33.
Any ideas?
I'm pretty sure contractions are not treated as words.
Instead the apostrophe is treated as a word separator giving you "I" and "m".
But you probably don't have innodb_ft_min_token_size=1, so those two "words" are ignored.
There are limitations of FT; you have encountered one of them.

How to restrict entire FTS5 Query to a single column?

Currently, I'm trying to execute an FTS5 query via libsqlite, and need to restrict the query to a specific column. In FTS4, this was possible by doing:
SELECT foo, bar FROM tableName WHERE columnName MATCH ?
and then binding the search string to the statement. However, with FTS5, the LHS of the MATCH operator must be the FTS table name itself, and the column name must be a part of the query:
SELECT foo, bar FROM tableName WHERE tableName MATCH 'columnName:' || ?.
This works when the binded string is a single phrase. However, consider the search text this is great. The query then becomes:
SELECT foo, bar FROM tableName WHERE tableName MATCH 'columnName:pizza is great';
Only pizza is restricted to to the columnName, but the rest of the phrase is matched against all columns.
How can I work around this?
The documentation says:
A single phrase … may be restricted to matching text within a specified column of the FTS table by prefixing it with the column name followed by a colon character.
So the column name applies only to a single phrase.
If you have three phrases, you need to specify the column name three times:
tableName MATCH 'columnName:pizza columnName:is columnName:great'

sqlite3 with FTS4 table: Query returns wrong row

I have a weird issue with my FTS4 index in SQLite3, namely that a MATCH query for one term returns not the exact match but another, similar one, and vice versa.
Here is an example to illustrate it better:
SELECT name FROM test_idx WHERE name MATCH 'lehmbruck-museum';
-- "Lehmbruck-Archiv"
SELECT name FROM test_idx WHERE name MATCH 'lehmbruck-archiv';
-- "Lehmbruck-Museum"
It seems to have something to do with the dash, here is a similar case that exhibits the same behavior:
SELECT name FROM test_idx WHERE name MATCH 'some-thing';
-- "some-thang"
SELECT name FROM test_idx WHERE name MATCH 'some-thang';
-- "some-thing"
Here is how this test database is built, in case somebody wants to have a go at reproducing it:
CREATE VIRTUAL TABLE test_idx USING fts4(name);
INSERT INTO test_idx (name) VALUES
('some-thing'), ('some-thang'),
('Lehmbruck-Museum'), ('Lehmbruck-Archiv');
SELECT name FROM test_idx WHERE name MATCH 'lehmbruck-museum';
What you pass to MATCH here is a full text query expression. The - character is a unary operator in that expression language that is a stand in for the NOT set operation, and is certainly giving you your unexpected results. Notably - the exact opposite of what you expect! Of course, it is finding exactly what the query is instructed to find - the string lehmbruck and NOT museum at the end!
You'll need to escape it to get the results you want - or perhaps employ the LIKE operator if you are looking at a single column in a table.
Some more information on this expression language can be found in section 3 of the FTS3 and FTS4 documentation on the SQLite doc site here.