Understanding `regexp_match()` in QGIS

The regexp_match() function searches a string using a regular expression (regex) and returns the starting position of the first match. If the pattern is not found, it returns 0.

This often causes confusion because when the function is used in tools such as Extract by Expression or Select by Expression, it appears to behave like a Boolean function. In reality, it always returns an integer, but QGIS interprets non-zero values as TRUE and 0 as FALSE when a Boolean result is expected.

Syntax

regexp_match(input_string, regex)

Argument	Description
input_string	The string to search.
regex	The regular expression pattern. Backslashes must be escaped in QGIS expressions (for example \\d, \\s, \\b).

Return Value

The function returns:

• 1 or greater — the position where the first match begins.
• 0 — no match was found.

For example,

regexp_match('Hello World', 'World')

returns

7

because World starts at the seventh character.

If the text does not exist,

regexp_match('Hello World', 'Earth')

returns

0

Why Does It Work in Extract by Expression?

Suppose the expression is

regexp_match("ref", '^NH')

and the attribute values are:

ref	Result
NH-01	1
NH45	1
SH-09	0
ABC123	0

The Extract by Expression algorithm expects a Boolean value.

Internally, QGIS converts the results as follows:

Returned value	Boolean
0	FALSE
Any non-zero number	TRUE

Therefore,

regexp_match("ref", '^NH')

behaves exactly like

regexp_match("ref", '^NH') > 0

or

regexp_match("ref", '^NH') = 1

because the pattern ^NH can only start at the first character.

Practical Challenge

Assume the karnataka_major_roads layer contains a field named "name" with the following values:

Road Name	Should Match?
1st Main Road	✓
2nd Main Road	✓
23rd Main Road	✓
101st Main Road	✓
Banneghatta Main Road	✗
Old Main Road	✗
Main Road	✗

The objective is to extract only roads that begin with a numbered ordinal followed by Main Road.

Step 1 — Identify the Pattern

Every valid road follows the same structure:

<number><ordinal> Main Road

Examples:

1st Main Road
2nd Main Road
23rd Main Road
101st Main Road

Step 2 — Build the Regular Expression

Break the pattern into smaller pieces.

One or More Digits

Regex:

\d+

In a QGIS expression:

\\d+

Ordinal Suffix

Possible suffixes are:

st
nd
rd
th

Regex:

(st|nd|rd|th)

The vertical bar (|) means OR.

A Space

Regex:

\s

QGIS expression:

\\s

Literal Text

Main Road

Beginning of the String

To ensure the road name starts with the number:

^

End of the String

To prevent matches such as 1st Main Road Extension:

$

Complete Regular Expression

Standard regex:

^\d+(st|nd|rd|th)\sMain Road$

QGIS expression:

'^\\d+(st|nd|rd|th)\\sMain Road$'

Using regexp_match()

regexp_match(
    "name",
    '^\\d+(st|nd|rd|th)\\sMain Road$'
)

What Does It Return?

Road Name	Result
1st Main Road	1
2nd Main Road	1
23rd Main Road	1
101st Main Road	1
Banneghatta Main Road	0
Main Road	0

Notice that every valid match begins at the first character, so the function returns 1.

Understanding the Return Value

Consider the road name:

23rd Main Road

The regular expression starts matching immediately:

2 3 r d   M a i n   R o a d
^

The match begins at the first character.

Result:

1

Now consider:

Banneghatta Main Road

The regular expression expects

^\d+

which means the string must begin with one or more digits.

Instead, the first character is B, so no match exists.

Result:

0

Another Example

Suppose the regular expression is simply

'Main Road'

Now the function searches anywhere in the string.

Road Name	Result
1st Main Road	5
23rd Main Road	6
Banneghatta Main Road	13

For "1st Main Road":

1 s t _ M a i n _ R o a d
1 2 3 4 5

The substring Main Road begins at character 5, so

regexp_match("name", 'Main Road')

returns

5

This demonstrates the real purpose of regexp_match(): it tells you where the first match occurs, not simply whether it exists.

regexp_match() vs LIKE

A LIKE expression such as

"name" LIKE '%Main Road'

would also match

• Banneghatta Main Road
• Old Main Road
• East Main Road

It cannot enforce the rule that the name must start with a numbered ordinal.

The regular expression

regexp_match(
    "name",
    '^\\d+(st|nd|rd|th)\\sMain Road$'
)

matches only names that satisfy all of the following conditions:

• begin with one or more digits;
• are immediately followed by st, nd, rd, or th;
• contain a single space;
• end with the exact text Main Road.

Key Points

• regexp_match() returns the starting position of the first regex match.
• If no match exists, it returns 0.
• In Boolean contexts, 0 is treated as FALSE and any non-zero value as TRUE.
• Anchors such as ^ and $ make it possible to validate the entire string instead of searching for a substring.
• Regular expressions are significantly more powerful than LIKE when matching structured text such as road names, parcel identifiers, postal codes, or formatted reference numbers.