Hi Dmitry

Perfect example for a cursor! Create the cursor so it matches the adir array or use sys(2000) to find all the files. Dump them into the cursor. Make sure there's a field for the sort order and use replace to fill that in OR add an index on the padded numeric portion of the file name.

Piece of cake and no worries about redimensioning arrays etc.

>Hi Mike,
>My case if very simple. I have to retrieve a bunch of XML files from a folder on the disk and process them in a certain order (by file name + suffix).
>
>The file name format is as follows "DDMMYYYY_##.XML". The ## in the file name is as follows: "1", "2",... "10", "11", etc. So to process the files in the order by the ##, I have to pad the ## with "0". So what I will do, once I get the array of file names, create a new array and copy the file name from the original array to the newly created array, at the same time padding the ## with "0". Then I can ASORT() the new array and process the files in the correct order.
>
>
>>Hi Dmitry
>>
>>What are you using an array for? A cursor often can be used instead.
>>
>>>Hilmar,
>>>I actually am using your approach 2 (even though you marked it with "yech" <g>.)
>>>This is probably one thing that I miss from Clipper (which is how I thought arrays should work) as compared to VFP arrays.
>>>
>>>Thank you for your input.
>>>
>>>
>>>>
>>>>As explained in other replies you got, the array is, unfortunately, treated as a one-dimensional structure; the two dimensions are just a convenient way to access it. But in the case of redimensioning, you are in trouble.
>>>>
>>>>Some options:
>>>>
>>>>

Give the array sufficient columns in the first place.
>>>>

To redimension, copy to another array (yech!).
>>>>

Use my "large array" class from the download section, instead of normal arrays. Accessing the array elements is slower than with built-in arrays, but redimensioning is no problem. Also, you can have more than 2 dimensions in your array.