RE: Using %scan and %subst vs pointers

Absolutely agree.

Unless there's a hugely egregious penalty, make it easier for the next person who has to work on it - it might be you!


-----Original Message-----
From: RPG400-L <rpg400-l-bounces@xxxxxxxxxxxxxxxxxx> On Behalf Of Daniel Gross
Sent: Sunday, June 5, 2022 8:57 AM
To: RPG programming on IBM i <rpg400-l@xxxxxxxxxxxxxxxxxx>
Subject: Re: Using %scan and %subst vs pointers

Hi,

Am 05.06.2022 um 16:38 schrieb Vinay Gavankar <vinaygav@xxxxxxxxx>:

For a given string (about 300 bytes with 34 fields, which is expected to be
the average), I iterated each of them 10000 times and captured the time
taken by each:
#1 - 70 milli-seconds
#2 - 86 milli-seconds
#3 - 712 milli-seconds

Not sure if my code for #3 is not efficient, but did not expect it to be
that bad. Even though #2 has the simplest code, I would probably go with #1.

For what it's worth - I wouldn't go with 13 lines of code for 16/10000 milli-seconds if I can have 3 lines of really simply and obvious code.

If you think about the 30 minutes that the next guy is trying to understand the code vs. the 1 minute you can parse over 1 billion lines of data, before your code is more efficient ;-)

If the advantage is that small - go with the code that's easier to understand and maintain - the next-poor-guy will thank you.

Just my 2 ct

Regards,
Daniel