>> Search Key

[Jonathan]

Hi I'm trying to import a large text file into Access. With or without an
import specification I get the following 'Import Text Wizard' error message...

The search key was not found in any record.

What does this mean?

Many thanks,
Jonathan

 

[u'are'hell]

YOUR HELL

 

[John Nurick]

Hi Jonathan,

No one quite seems to know what this means. Possibilities seem to
include:

1) An irregularity in the structure of the text file.

2) Corruption in the database you're importing into.

3) The amount of data to be imported takes the .mdb file over the 2GB
limit (even if the initial size of the .mdb file plus text file is
less than 2GB, space is needed for indexes and maybe temporary
structures).

(1) seems to be behind a lot of problems importing big text files. One
approach is to split the file in two and try importing the halves
separately. The half that won't import contains the problem record, so
split that in two and try again. Eventually this binary approach gets
everything imported except the problem record which can be
reconstructed manually.

For (3), it helps to compact the .mdb before importing. If necessary,
delete as many as possible of the indexes on the table in question,
import the data, then re-index.

 

[Jonathan]

Hi John, thanks for this detailed response as this is going to be a repeated
process I hope that it is solved by a simple compact...

Many thanks, Jonathan

 

[John Nurick]

Hi Jonathan,

"Large text file" may not mean the same thing to you and to me. If
we're talking tens or hundreds of megabytes there are a few things
worth thinking about, including:

* Often, one only needs to import some of the columns and/or some of
the records; the rest are of no interest. In this case one can link to
the text file and then use an append query to select the data needed
and move it into an Access table.

* If it's just a matter of analysing and reporting on the data in the
text file, often it can be done by linking to the file and querying
the linked table without importing the data at all.





* most big t