>>>>>>Hi,
>>>>>>
>>>>>>The way to populate an array from a current cursor is as follows:
>>>>>>

>>>>>>local array aDbfStruct[1]
>>>>>>afields( aDbfStruct )
>>>>>>

>>>>>>
>>>>>>But if you are using a SQL Server Native Client, some fields (e.g. Varchar(MAX) fields) will have the length 0.
>>>>>>So, I need to remove these elements from the array aDbfStruct. Or exclude these elements when creating another cursor using the code:
>>>>>>

>>>>>>create cursor CUR_MYCURSOR from array aDbfStruct
>>>>>>

>>>>>>
>>>>>>How do you remove all elements from the above array which have element 3 equal to 0 (zero)? Or, can the above CREATE CURSOR be modified to ignore the "bad" fields?
>>>>>>
>>>>>>UPDATE: I can scan the array and check each element 3 for a value of 0 and delete it. But I thought maybe there is a more efficient way to delete those elements.
>>>>>>
>>>>>>TIA
>>>>>like this?
>>>>>

>>>>>COPY STRUCTURE EXTENDED to TabXYZ
>>>>>sele cur1.* from TabXYZ as Cur1 into array aDbfStruct where FIELD_LEN=0
>>>>>use in TabXYZ
>>>>>dele TabXYZ.*
>>>>>

>>>>>
>>>>>and
>>>>>

>>>>>COPY STRUCTURE EXTENDED to TabXYZ
>>>>>dele from TabXYZ where FIELD_LEN=0
>>>>>create newtable from TabXYZ
>>>>>use in TabXYZ
>>>>>dele TabXYZ.*
>>>>>

>>>>>should work w/o array
>>>>
>>>>I will try this. Thank you. Although I am thinking if this is more efficient that deleting elements from the array.
>>>
>>>It should be better then looping through the array
>>>No FOR .. ALEN() to 1 step -1, no IF, no ADEL, no DIMENSION.
>>>Faster, less complex code.
>>>The need to create a file and delete it is odd, but that can't be helped. Try to target GETENV('temp') dir
>>>update: consider to use some unique name like sys(2015) too. Much fun if two process create the same file .... SYS(2015) is not fool prove, but less likely.
>>
>>With my luck, the less likely will turn out to be :) I will consider your approach. Thank you.
>
>Then you must create a UUID and name the file like that. This is fool prove. (But, given the extra time to create the UUID a FOR NEXT loop might be finished in this time too) ;)

I think I will stick with scanning the array and deleting the "bad" columns. One approach vs the other could be a few millisecond difference. The customer will have to wait :)