this stage is running from 45 mins, did i do somet...
# suitescript
r
this stage is running from 45 mins, did i do something wrong, i dont even have summarize stage..
n
🤷‍♂️ how could anyone know except you?
s
this is the universe telling you to use Scheduled scripts instead.
unless of course you actually need MR technology. Then the tradeoff is less reliable ops.
a
This is the first time I have ever heard that. Isn't the usual advice to always use MR over scheduled?
n
Governance requirement is likely going to drive your choice. In SS1.0 you can yield a scheduled script in SS2.0 you cannot and have to get creative.
s
No, it's a myth than MR is supposed to be used over Scheduled scripts. At least I've never seen such an assertion - I think it's more a matter of developers and the shiny new thing effect.
Also, I'd never use yield in a script because it was unreliable.
IMHO, scheduled scripts remain the correct tool for many jobs.
e
Is this the same philosophy for using saved searches instead of suiteanalytics workbooks?
b
with the implication being that scheduled scripts are inappropriate for processing large amounts of data
the usual interpretation being anything requiring more than 10k points
s
I reach for MR if there is "large" amounts of data and the logic on that data fits well into a parallel model
a
I have never used scheduled scripts, but what's the downside of a MR or the advantage of a scheduled script?
n
map/reduce can't use the script debugger
but can operate over larger datasets
a
Scheduled scripts can use script debugger?
s
Scheduled scripts are a simpler model, ideally suited for serial operations that must happen in a specific order.
they are also easier to debug (even without the debugger) and you have full control whereas in MR you have less control (in return for parallel execution)
If you're not taking advantage of parallel execution my impression is that scheduled scripts also run faster than MR