Regular expressions

[Markus Innerebner]

Hello to everyone,

Yesterday I tried long time to make a validation for a number format
input field.
As I am using Regex in Java I wrote following expression:

String pattern =
"((\\d+)((\\.)(\\d+)(\\,(\\d)+)?)?)|((\\d+)((\\,)( \\d+)(\\.(\\d)+)?)?)";

So the sequence of the chars of the input field could be?

123 or
123.89 or
123.89,99 or
123,89
123,99.88

I used the function search() but I wasn't able to get the same result as
I get in Java.

Any Hints?
Thanks

Markus

[Evertjan.]

Markus Innerebner wrote on 24 mrt 2006 in comp.lang.javascript:

Hello to everyone,

Yesterday I tried long time to make a validation for a number format
input field.
As I am using Regex in Java I wrote following expression:

String pattern =
"((\\d+)((\\.)(\\d+)(\\,(\\d)+)?)?)|((\\d+)((\\,)( \\d+)(\\.(\\d)+)?)?)";

So the sequence of the chars of the input field could be?

123 or
123.89 or
123.89,99 or
123,89
123,99.88

I used the function search() but I wasn't able to get the same result as
I get in Java.

Any Hints?

Better ask a Java NG, this NG is about Javascript.

If you are talking javascript show your code.
--
Evertjan.
The Netherlands.
(Please change the x'es to dots in my emailaddress)

[Vic Sowers]

"Markus Innerebner" <ma****@innerebner.net> wrote in message
news:44***********************@reader2.news.tin.it ...

Hello to everyone,

Yesterday I tried long time to make a validation for a number format input
field.
As I am using Regex in Java I wrote following expression:

String pattern =
"((\\d+)((\\.)(\\d+)(\\,(\\d)+)?)?)|((\\d+)((\\,)( \\d+)(\\.(\\d)+)?)?)";

So the sequence of the chars of the input field could be?

123 or
123.89 or
123.89,99 or
123,89
123,99.88

I used the function search() but I wasn't able to get the same result as I
get in Java.

Any Hints?
Thanks

Markus

I can't decipher yours (just lazy) but I use:

/^((\d{1,3}(,\d{0,3})*)|\d+)(\.\d*)?$/

in JavaScript.

Vic

[Lasse Reichstein Nielsen]

Markus Innerebner <ma****@innerebner.net> writes:

Yesterday I tried long time to make a validation for a number format
input field.
What are the valid formats? Examples rarely cover the entire range
of valid values.
As I am using Regex in Java I wrote following expression:

String pattern =
"((\\d+)((\\.)(\\d+)(\\,(\\d)+)?)?)|((\\d+)((\\,)( \\d+)(\\.(\\d)+)?)?)";
Ick. Completely unreadable. Let's try to convert it to a RegExp literal
and see if it can be reduced further. I can see "\\," where the escape
isn't necessay:

/((\d+)((\.)(\d+)(,(\d)+)?)?)|((\d+(,)(\d+)(\.(\d)+ )?)?)/

So it's two parts:
((\d+)((\.)(\d+)(,(\d)+)?)?)
and
((\d+(,)(\d+)(\.(\d)+)?)?)

It sounds unlikely that you want to capture all these sub-parts. Or at
least, the (\d)+ should probably be (\d+) if you want to use it for
something.

Let's remove unnecessary brackets - they're only necessary when you
use it, not for unerstanding :). That gives:
\d+(\.\d+(,\d+)?)?
and
(\d+,\d+(\.\d+)?)?

I.e.
1) some digits optionally followed by (a period and some digits,
optionally followed by a comma and some digits)
or
2) nothing, or some digits a comma and some digits, optionally
followed by a dot and some digits.

So, it appears that the format is:
zero or more digits, optionally separated by a comma and/or a period,
i.e., comma and period cannot be at the ends or next to each other.
Is this correct?
So the sequence of the chars of the input field could be?

123 or
123.89 or
123.89,99 or
123,89
123,99.88

I used the function search() but I wasn't able to get the same result
as I get in Java.
What results did you get here, and what results do you get in Java?

Notice that your RegExp isn't anchored, so the default is to test
whether it can match any substring of the tested string. Since it
acceptes the empty string, this is always the case.
Any Hints?

Specify, very precisely, which strings should be accepted.

Consider whether doing it all in one RegExp is better than
doing it in more steps, or in code, e.g.:

function validateString(string) {
if (string == "") { return true; }
if (/,.*,/.test(string)) { return false; }
if (/\..*\./.test(string)) { return false; }
return /^\d+([.,]\d+){0,2}$/.test(string);
}
/L
--
Lasse Reichstein Nielsen - lr*@hotpop.com
DHTML Death Colors: <URL:http://www.infimum.dk/HTML/rasterTriangleDOM.html>
'Faith without judgement merely degrades the spirit divine.'

[Markus Innerebner]

Hello Lasse (and to the other repliers),
First of all thanks for your answers.

/((\d+)((\.)(\d+)(,(\d)+)?)?)|((\d+(,)(\d+)(\.(\d)+ )?)?)/
this is that what i did at the begin. It's better to remove the brackets
and split it in two parts, because it's better readable
So it's two parts:
((\d+)((\.)(\d+)(,(\d)+)?)?)
and
((\d+(,)(\d+)(\.(\d)+)?)?)

It sounds unlikely that you want to capture all these sub-parts. Or at
least, the (\d)+ should probably be (\d+) if you want to use it for
something.

Let's remove unnecessary brackets - they're only necessary when you
use it, not for unerstanding :). That gives:
\d+(\.\d+(,\d+)?)?
and
(\d+,\d+(\.\d+)?)?

I.e.
1) some digits optionally followed by (a period and some digits,
optionally followed by a comma and some digits)
or
2) nothing, or some digits a comma and some digits, optionally
followed by a dot and some digits.

So, it appears that the format is:
zero or more digits, optionally separated by a comma and/or a period,
i.e., comma and period cannot be at the ends or next to each other.
Is this correct? This is exactly that what i want. What was missing here (and in my code
too) are the symbols ^ at the begin and $ at the end, so that other
characters are excluded.

Specify, very precisely, which strings should be accepted.

Consider whether doing it all in one RegExp is better than
doing it in more steps, or in code, e.g.:

function validateString(string) {
if (string == "") { return true; }
if (/,.*,/.test(string)) { return false; }
if (/\..*\./.test(string)) { return false; }
return /^\d+([.,]\d+){0,2}$/.test(string);
}

This is a good idea too! Probally I will use something like that.
Now it's clear for me that r.e. in Java are differently than r.e. in
javascript (or perl).
I forget to set the chars '^' on the begin and '$' at the end.

example:

in java: \\d+ can be: 1, 123, but not 12abc
in javascript: /\d+/ can be: 1, 12, and also 12abc, just look if
the string contains a number
so I changed to: /^\d+$/ and this r.e. returns the same result as my
r.e. of java.

Anyway, thanks for helping be

best regards

Markus

[Thomas 'PointedEars' Lahn]

Markus Innerebner wrote:

Now it's clear for me that r.e. in Java are differently than r.e. in
javascript (or perl).
They are different, but in a different way than you think.
I forget to set the chars '^' on the begin and '$' at the end.

example:

in java: \\d+ can be: 1, 123, but not 12abc
You have to escape the backslash (`\') in the Java String literal,
because it starts an escape sequence. This is not different in
ECMAScript-conforming String literals.
in javascript: /\d+/ can be: 1, 12, and also 12abc, just look if
the string contains a number

If you used `new RegExp("\\d+")' in an ECMAScript implementation, you could
use the Java String literal verbatim. The issue here is instead that Java
does not appear to have a Regular Expression literal syntax, while
implementations of ECMAScript, such as JavaScript and JScript, have one.

Using RegExp literal syntax avoids escaping the backslash again and makes
Regular Expressions easier readable. Another advantage is that the RegExp
object is created only once per expression because it is created before
execution. However, it also introduces a dependency on JavaScript 1.2 (NN
4.0, June 1997 CE), JScript 3.0 (IE 3.0, October 1997 CE), or an ECMAScript
3 (December 1999 CE) implementation, and has the drawback that it cannot
span lines, which is possible with the RegExp() constructor (through
string concatenation or joining of an array of string values).
PointedEars