From: Nobody on
"C" <wrong.address.1(a)gmail.com> wrote in message
news:c5ff7d34-964d-4cfd-9a66-
> I am inclined to
> learning for another few months with trial and error before starting
> recoding.

F2 is your friend then. It brings up Object Browser, which is the quickest
way to learn .Net framework. Try searching for "graphics" for instance.

Also, download MSDN Library so you have full reference. Here is the download
link:

http://www.microsoft.com/downloads/details.aspx?displaylang=en&FamilyID=7bbe5eda-5062-4ebb-83c7-d3c5ff92a373

You need to uninstall the Express version of the library before installing
the full version as the second note in the link above says.


From: Alex Clark on
Mike,

I'd be happy to take that challenge. Case study:

Recordset with 10,000+ records which need to be written out to a file in CSV
format. Fairly standard, run of the mill export operation that you find in
loads of LOB apps these days. Oh, and we're working with strings so VB6
should be far quicker than .NET... right?

Let's see what you can do, and I guarantee .NET will better it in terms of
performance.

Alex


"Mike Williams" <Mike(a)WhiskyAndCoke.com> wrote in message
news:O6pUQEqPKHA.1796(a)TK2MSFTNGP02.phx.gbl...
> "Cor Ligthert[MVP]" <Notmyfirstname(a)planet.nl> wrote in message
> news:%23BuPb2pPKHA.1280(a)TK2MSFTNGP04.phx.gbl...
>
>> Despite what others say, if you want at the end a
>> program that is 10 times faster then VB6 (what is
>> possible with VB for Net) . . .
>
> Put your money where your mouth is fatty. And change that picture on your
> MVP profile. You look like a child molester with his head on upside down.
>
>
>
>


From: Schmidt on

"Alex Clark" <quanta(a)noemail.noemail> schrieb im Newsbeitrag
news:O3K1IwyPKHA.4428(a)TK2MSFTNGP02.phx.gbl...


> I'd be happy to take that challenge. Case study:
>
> Recordset with 10,000+ records which need to be written out
> to a file in CSV format. Fairly standard, ...

Ok, in case you want to play with FileIO and String-Handling
(parsing, concatenating) - I'd be willing to take up this
"challenge", in case we can leave out the "DB-stuff", since
that would make the examples easier to implement on
both sides.

What, if the "Raw-Data" does not come from a DB, but from
a 'Src.csv'-File instead?

The task would then be, to read-in (parse) the 'Src.csv',
order the records by a certain String-Field and write the
sorted results back to a 'Dst.csv'.

Timing would then be measured "over-all" (parsing the Src.csv,
sorting, writing back to Dst.csv).

Here comes a concrete task-proposal:
Download 'holdings.csv' from:
http://www.archive.org/details/LibraryHoldingsData
which is ~8MB gz-compressed and unpacks to an about
80MB csv-File with "non-quoted string-field-content".

It contains only 4 Fields per record:
ISBN (String)
CarnegieNr (Integer)
LibName (String)
LibURL (String)

So, let's just parse that in, perform a simple binary sort (or
if you want, then also a caseinsensitive one) on the third
Field: LibName (String) - and write the sorted records
back to a Dst.csv in ascending (LibName) order.
To prevent "record-copy-over-trickery" at the ByteLevel,
there should be an additional condition, to write out each
String-Field content of the new File within additional
csv-quotes ("...").

That's an easy enough task-description, dealing with Files
and Strings, leaving out any ADO/ADO.NET- and DB-
dependencies.

If you agree about the outlined task, then we could start
immediately with that.

Olaf