I replaced a python/mysql daily process that takes 25 minutes to run with a perl/redis process that takes 1 minute to run, so it runs multiple times a day. Mgmt asks me to convert it to python/mysql, "...but keep the run time at one minute. That's great!"

No.

If works faster not need to reimplement imho.

depends on implementation, but I have yet to see any application that is faster on nosql database than on a well optimized relational database.

I also doubt it has anything to do with py/perl. Probably just a different implementation.

Possibly as well, MySQL is kinda shit, but that alone can't account for 25x difference.

@AndSoWeCode

Sure. The app I replaced was very poorly written, and supported by a bad programmer. The approach it takes is bad, and a ground-up rebuild was the only solution I saw as workable.

That said, nothing outruns perl in text processing, except compiled code, and no backend operates faster than redis.

@bahua yeah, but differences in performance Perl vs Python in text performance are what? 25%?

Also, I can easily give you several scenarios where Redis would be a slow implementation for a backend. And can probably optimize a good relational database to perform at speeds close to what you have on Redis, or even more.

I would also probably fuck up a Redis implementation.

You know why? Experience and ability. That code was slow because of the developer. My redis implementation would be slow and shitty because I'm a newb, your relational DB implementation will be worse than on Redis because you probably have more experience with Redis.