On Wed, 29 Dec 2021, Graeme Geldenhuys via fpc-pascal wrote:

On 2021-12-29 09:54, Michael Van Canneyt via fpc-pascal wrote:
They showed clearly that the used protocol (http & xml) is the largest factor. in 2013 I did many tests in a JSON REST framework: the times for streaming were dwarfed by the time spent on network & database access.

Have you, or anybody else, ever done some benchmarks to see how "expensive" in CPU time it is to parse the exact same data in XML format vs JSON format?

I couldn't find FPC specific benchmarks, but did a basic Firefox v95.0.2 browser Network tests (with caching disabled). A 54KB JSON file vs a 121KB XML file, which represents the exact same 1000 rows of a database table. Interestingly (and surprising to me), Firefox managed to load the XML in about half the time (172ms) compared to the JSON (386ms) file. I did the load test multiple times, and the results were consistent.

It depends on how you load it. It's quite possible the XML uses a SAX model
where the nodes are not loaded till you actually access them.
So I would definitely also add a loop that runs over all rows.

I did a test once on pure FPC, and there the JSON worked faster.

Michael.
_______________________________________________
fpc-pascal maillist  -  fpc-pascal@lists.freepascal.org
https://lists.freepascal.org/cgi-bin/mailman/listinfo/fpc-pascal

Reply via email to