Comparing String Lengths with Different Replacements in PHP

Vladimir A. · 2024-03-01T17:32:11.688Z

who can tell what does this do: strlen(replace(?text,"[\\p{L}0-9]+[^\\p{L}0-9]*","1")) And compare it against strlen(replace(?text,"[\\p{L}0-9]+[^\\p{L}0-9]*","2"))

The Knowledge Graph Conference

9 comments

· Sorted by Oldest

Wolfgang S.
·
My guess would be: Regular expression (see https://docs.oracle.com/en/java/javase/11/docs/api/java.base/java/util/regex/Pattern.html#ucc) \\p{L} any unicode character 0-9 a digit in the range 0-9 + at least one * zero or more [....] any of the characters mentioned in the brackets [^...] anything but the characters mentioned in the brackets so [\\p{L}0-9]+[^\\p{L}0-9]* should match a text starting with at least one unicode character or a digit followed by zero or more symbols which do not fall in that category replace (see https://www.w3.org/TR/sparql11-query/#func-replace) any part in the provided ?text matching the expression above will be replaced by the digit 1 (or 2 in the second variant) The overall result is the string length of the result of that replace operation. You might have reached that same conclusion already, so not sure if I could provide any help in interpreting?
Vladimir A.
·
So, what is the purpose of it?
Wolfgang S.
·
No idea, could be to "normalize" some identifier to a single digit string to determine how many there are in a source value? Do you have an example value?
Vladimir A.
·
"Is foo+bar = baz?" -> 4
Wolfgang S.
·
maybe kind of a "word count"?
👍1
Wolfgang S.
·
the two different occurences should not return different values. Essentially it shouldn't matter what you replace it with, i.e. whether to use 1, 2, or x as replacement value, if the only purpose is counting the length of the resulting string
Vladimir A.
·
This for Ontotext KG, to populate for articles: s:wordCount and s:timeRequired (at 284 wpm, assumed rate)
Phil T.
·
Wolfgang S.’s assumption appears correct. [\\p{L}0-9]+ means at least one letter from any natural language code-set (i.e. alphabet) or number in any order. So, foo, fo9o, foo9, and 9foo will match. [^\\p{L}0-9]* means anything else (whitespace, symbols) and is optional. So the pairing produces a single match for ‘foo $/*. ’ and replace will return ‘1’ Thus, ‘Thé quick br0wn føx leaped 9 times.’ would produce ‘1111111’ whose string length is 7. So yes, appears to be a word count that supports world languages.
Marcelo Barbieri
·
Can try it here with some test data, too 😉 https://regex101.com
👍1

Wolfgang S.
·
My guess would be: Regular expression (see https://docs.oracle.com/en/java/javase/11/docs/api/java.base/java/util/regex/Pattern.html#ucc) \\p{L} any unicode character 0-9 a digit in the range 0-9 + at least one * zero or more [....] any of the characters mentioned in the brackets [^...] anything but the characters mentioned in the brackets so [\\p{L}0-9]+[^\\p{L}0-9]* should match a text starting with at least one unicode character or a digit followed by zero or more symbols which do not fall in that category replace (see https://www.w3.org/TR/sparql11-query/#func-replace) any part in the provided ?text matching the expression above will be replaced by the digit 1 (or 2 in the second variant) The overall result is the string length of the result of that replace operation. You might have reached that same conclusion already, so not sure if I could provide any help in interpreting?
Vladimir A.
·
So, what is the purpose of it?
Wolfgang S.
·
No idea, could be to "normalize" some identifier to a single digit string to determine how many there are in a source value? Do you have an example value?
Vladimir A.
·
"Is foo+bar = baz?" -> 4
Wolfgang S.
·
maybe kind of a "word count"?
👍1
Wolfgang S.
·
the two different occurences should not return different values. Essentially it shouldn't matter what you replace it with, i.e. whether to use 1, 2, or x as replacement value, if the only purpose is counting the length of the resulting string
Vladimir A.
·
This for Ontotext KG, to populate for articles: s:wordCount and s:timeRequired (at 284 wpm, assumed rate)
Phil T.
·
Wolfgang S.’s assumption appears correct. [\\p{L}0-9]+ means at least one letter from any natural language code-set (i.e. alphabet) or number in any order. So, foo, fo9o, foo9, and 9foo will match. [^\\p{L}0-9]* means anything else (whitespace, symbols) and is optional. So the pairing produces a single match for ‘foo $/*. ’ and replace will return ‘1’ Thus, ‘Thé quick br0wn føx leaped 9 times.’ would produce ‘1111111’ whose string length is 7. So yes, appears to be a word count that supports world languages.
Marcelo Barbieri
·
Can try it here with some test data, too 😉 https://regex101.com
👍1