Hi Tore

Hmmm, sounds like a load of faff to me. The problems are:

1) I don't know WHAT columns will be included in the input file.
2) I don't know how many columns
3) I don't know, therefore, what data type they'll be
4) Even if I import to all character fields, who's going to sort out the mapping and data typing (remember that this is a user job) in the analysis?
5) million other problemettes :-)


>Create a dummy cursor, and append the data into it. Analyze the data in cursor, and then create a new cursor with the correct structure.
>
>>Hi
>>
>>This is not something that I can't do, after a lot of faff, but I thought there may be a better, simpler idea out there, so I'm looking for suggestions/solution.
>>
>>I have an Ordnance Survey PointX text file of geographical entities, such as post offices, doctors' surgeries, etc., to import, including lat/long values. It's in the form of a delimited file, e.g.
>>
>>Record Type|Unique Reference Number|TOID|TOID Version|Name|...
>>"L"|14357552|"1000000347714764"|3|"Constitutional Hill"|...
>>
>>I did the software long ago. I created an intermediate file, with fields to match the column headings in the PointX file and imported it via:
>>
>>Append from ( .cFacsDataFile) Delimited with " with character |
>>
>>All well and good but, originally, there was a telephone No. column. I've just imported some more data and found that my longtitude value was all to cock. This is because in the latest import file there is no tel. # column, so they're all "shifted along" 1
>>
>>Now the above append command is nice and easy and I don't want to have to write a shed load of new code to get around changes in the format of the input data.
>>
>>Any ideas out there, please?
>>
>>'ppreciate it
>>
>>Terry