I have a problem that sounds very simple at first:

I need to match single digits in a string, but leave out sequences made up of more than one consecutive digits.

The issue is: I build an online dictionary using php/mysql. I use html encoding for the dictionary. Examples:

(First the string stored in database then how it should look on the webpage)

autorizált kereskedő (shown on the webpage as autorizált kereskedő)

SZISZ1 (shown on the webpage as SZISZ¹)

SZISZ2 I. (shown on the webpage as SZISZ² I.)

I need to match the 1 and 2 from the words "SZISZ1" and "SZISZ1 I." to convert them to superscript  but leave out the sequence 337 from the html encoding of special characters in other words. All special characters have a similar code in the database: "&#[0-9]{3}" and all single digits have the form "[0-9]{1}". My problem is that the regular expression used to find a single digit fires separately for each digit of a sequence of digits and it messes up how my special characters shown on screen.

Please do not tell me to store the superscript html tags in the database.

Your help will be greatly appreciated.

Thank you, Tihamer

pls be consistent with your specification:

you wrote: 

***I need to match the 1 and 2 from the words "SZISZ1" and "SZISZ1 I." to convert them to superscript  but leave out the sequence 337 from the html encoding of special characters in other words.***

 the problem is there is no string *337* in the two inputs you gave: "SZISZ1" and "SZISZ1 I."

keep your spec simple, like

i got string

SZISZ1__337XXXi need to only match on

SZISZ1

if you give an extra effort to explain it better, you'll most likely get help faster here.

Sorry, I try to be explicit now:

I have a list of words (an online dictionary).

Examples: 

szisz1,

szisz2

szisz1 II.

szisztĸm

I need a regular expression to match and replace the numbers 1, 2 and 1 from the first three words, but not to do anything to the 312 in the fourth word.

Result what I need:

szisz¹

szisz²

szisz¹ II.

szisteĸm (no change here)

Thank you.

Found a solution:

Search pattern: '/([a-zA-Z])(\d)/'

Replacement pattern: '$1^$2'

That is: I search for a sequence made up of one letter [a-zA-Z] plus one digit \d

and in the replacement I keep the letter ($1) (use it in the search pattern for context only) and add the  around my digit ($2).

This is basically scratching my left ear with my right hand and the problem still remains:

Can you use a regular expression to match ONE AND ONLY ONE digit in a string?

Try (untested):

(?<!\d)\d(?!\d)

Susan

Thank you for your advice, Susan.

It does not seem to work, but I got the idea, and it can be refined if tested.

Tiha

In what way does it "...not seem to work..."? Does it give you syntax or other errors - if so what are they? Does it find too much or not enough or what?

It might help if you read the posting guidelines (in the sticky note at the beginning of the forum) and told us the regex variant you are using.

Susan