Slow indexing of DBF database

[Greg Stachowski]

Hi,

I have a database of 215,000 records. The legacy Clipper for DOS application indexes it in a couple of seconds. Python with dbf takes about 45 seconds suing create_index. Is there any way to speed this up?

Regards

Greg

[Ethan Furman]

On 07/08/2015 12:47 PM, Greg Stachowski wrote:

> I have a database of 215,000 records. The legacy Clipper for DOS application indexes it in a couple of seconds. Python with dbf takes about 45 seconds suing create_index. Is there any way to speed this up?

How many fields are you using?

What types are they?

What does your index expression look like?

--
~Ethan~

[Greg Stachowski]

On Wednesday, 8 July 2015 21:57:05 UTC+2, Ethan Furman wrote:

How many fields are you using?

What types are they?

What does your index expression look like?

- indexing over 2 fields

- both strings (5 and 3 characters long)

- g.create_index(lambda rec: (rec.a, rec.b))

--
Greg

[Ethan Furman]

The only hope you have of a faster index is using a newer version (assuming you are using an older one).

Adding .idx files is in the works, but it's not done yet.

--
~Ethan~

[Greg Stachowski]

I'm using 0.96.003.

Ok. It's not too much of an issue. It affects startup time, but I can live with that. Once the index is done the rest is pretty quick. Thanks for the quick response and the work you've done on dbf.

Greg