Splitting A Line Into Two or More Effectively

[Melvin Mah]

I have one line formed as a continuous String as in:

<arg3><instance>aInstance</instance><property>aProperty</property><value>sdsds</value><value>d333</value></arg3><#INS#><arg3><instance>bInstance</instance><property>bProperty</property><value>232ada</value></arg3>

.

I am not very sure what is the best way to split into pairs of <arg3>(line)</arg3>. I have experimented with using indexOf but noticed that it is not quite effectively. Someone also mentioned about using regex to split but I am not sure how effective can it be in using that for the long string above.

If it is, what is the right regex for that above?

Thanks.

[Jesper de Jong]

What exactly do you mean with "effective"?

What did you try (show us your code)? Did it do what you expected, or not? What exactly do you expect?

[Richard Tookey]

and what do you mean by 'split' ? i.e. what do you want to do with the text between the <arg3>(line)</arg3> chunks?

[Melvin Mah]

Ah okay. I think I wasn't clear enough.

Instead of one line as I am getting now, I want to break it to pairs (starting with <arg3>, ending with </arg3>). Each pair will be stored into a single String array.

The problem right now is that I am still not able to split it up.

[Richard Tookey]

And discard anything between? e.g. in </arg3><#INS#><arg3> to discard the <#INS#> ?

[Melvin Mah]

Richard Tookey wrote:And discard anything between? e.g. in </arg3><#INS#><arg3> to discard the <#INS#> ?

Yes. <#INS#> can be discarded. It's a mere separator.

[Richard Tookey]

I can immediately see two approaches -

1) Loop using indexof() first to find "<arg3>" and then to find "</arg3>" where the start point in each indexOf() is the last successful indexOf() result. Use substring to extract the part you want. Break the loop when either indexOf() fails.
2) Write a regular expression for split() that splits when one looks ahead to find "</arg3>" then reluctantly anything and then looks ahead to find ""<arg3>" .

I prefer the second option (it only takes one line) but if you are new to regular expressions you may find this difficult.

[Melvin Mah]

Richard Tookey wrote:I can immediately see two approaches -

1) Loop using indexof() first to find "<arg3>" and then to find "</arg3>" where the start point in each indexOf() is the last successful indexOf() result. Use substring to extract the part you want. Break the loop when either indexOf() fails.
2) Write a regular expression for split() that splits when one looks ahead to find "</arg3>" then reluctantly anything and then looks ahead to find ""<arg3>" .

I prefer the second option (it only takes one line) but if you are new to regular expressions you may find this difficult.

You are right. Second option is easier. I tried this ([^\\;\\]*[^;<]) and it's as close to what I'm looking at. For <#INS#>, I just substituted with a ";" an easier delimiter. Not sure it's good or not.

[Richard Tookey]

That doesn't look at all right.

[Jayesh A Lalwani]

Is the <INS#> seperator guaranetted to be there in your input stream? You could just seperate by the seperator...hence the name seperator