>>>>He everybody,
>>>>
>>>>Suppose, I have an array generated by afields(). I have DataDict table, which describes structure of a table, so I can generate similar array. What would be the quickest way to compare two such arrays?
>>>
>>>How about converting your AFIELDS() array to a cursor - you could then use SELECT - SQL against the two tables/cursors.
>>
>>It's still not clear, how then I compare two cursors... scan, scatter & compobj seems to be not easy.
>>
>>Thanks. I'll think about it.

* Get new fields not in DataDict:
>SELECT ;
>  * ;
>  FROM FieldsCursor ;
>  WHERE FieldsCursor.FieldName NOT IN ;
>    (SELECT FieldName FROM DataDict) ;
>  INTO CURSOR Cursor1
>
>* Get fields in DataDict no longer in the table:
>SELECT ;
>  * ;
>  FROM DataDict ;
>  WHERE DataDict.FieldName NOT IN ;
>    (SELECT FieldName FROM FieldsCursor) ;
>  INTO CURSOR Cursor2
>
>* If you massage the FieldsCursor to have the same structure as DataDict,
>* create a third cursor with the same structure:
>CREATE CURSOR Cursor3 ...
>
>* Append ALL fields from both source cursors:
>APPEND FROM FieldsCursor
>APPEND FROM DataDict
>
>* Use the GROUP BY clause on ALL the field attributes
>* e.g. type, size, decimals etc. If you get a COUNT of 2,
>* the fields are identical, if 1, there is a difference in one of the
>* field attributes:
>SELECT ;
>  FieldName, ;
>  COUNT(*) AS nCount ;
>  FROM Cursor3 ;
>  WHERE Cursor3.FieldName NOT IN ;
>    (SELECT FieldName FROM Cursor1) ;
>    AND Cursor3.FieldName NOT IN ;
>      (SELECT FieldName FROM Cursor2) ;
>  INTO Cursor4 ;
>  GROUP BY < list of all field attributes >
>  HAVING nCount < 2

Great! Thanks a lot, I'll probably then go forward with my idea.
I have a program modifystructure. It works just fine, though quite complicated. I was thinking about creating a custom object, which will have a_Fields array, a_Indexes and some other properties. Then I can easily manipulate with DataDict and oTable.

Thanks again.