so possible workaround ideas would be: • is there any way of running a map reduce in other role than administrator? (that could allow us to have multiple deployments one for each role) • is there a way of running a saved search under a specific role from code?

Yes, you can run as role on the m/r deployment. You have a few options how you set that. • Create a new deployment for each role you want to run as • Have the calling script edit the deployment with the role before executing If you can, I would create a new deployment for each role, maybe identify those deployments by the role ID so you know which one to call.

Hi, are you 100% sure? because when creating a deployment for map reduce, it does not let you select the role, it goes straight to administrator

Interesting. I don't think I ever realized that on the M/R script deployments. I know you can set it for other script types because I do this in one of my install scripts>

record.load({
                        type: 'scriptdeployment',
                        id: results[0].primarykey,
                        isDynamic: true
                    })
                    .setText({fieldId: 'runasrole', text: 'Administrator'})
                    .save();

yet another reason to not use M/R scripts 🙂

the problem is that we have like 2000 lines of code in that Map reduce, moving it to something else would be imposible

How about using a scheduled script to run the saved search and save the results to a file and then pass that fileId to the M/R script for processing?

we are running 3 saved searches on parallel and composing data based on it, it would be very difficult

So in your mr script are you saying that running the saved searches in parallel and combining data is done in the map or reduce stage?

Is getting the exchange rate and converting the currency values not an option?

@Eric B we are running a map reduce, we have a set of queries that are executed multiple times with different filters, both in the get and in the map

If you’re executing a set of queries multiple times using different filters in the getinput stage that wouldn’t run in parallel so I don’t see why you can’t use a scheduled script to do that part to retrieve the data you need. Architecturally speaking, you’re better off constructing your data object in the getinput stage and then grouping them in the map stage before you run your processing in the reduce stage.

I ran some tests and when you set a Suitelet to run as role it executes the search as that role. You would have to use N/HTTPS and set it was externally available which isn't ideal. I would suggest using a Restlet instead so the tokens used would be tied to the roles. That is the only solution I can think of short of what Eric mentioned with the search results being saved as a file.

You can create multiple deployment of the same map/reduce script and create a search script parameter (list/record, saved search) where you change the criteria of each search for each subsidiary currency. So basically each deployment of the map/reduce points to a different saved search (with different results based from the criteria) to process at the getInputStage by doing a search.load (instead of hard coding the search). Hope that make sense. Ex: customdeploy_mr_subA_usd —> transaction search (where all trx are from subA and usd) customseploy_mr_subB_cad —> transaction search (where all trx arefrom subB and cad)

you cannot specify the conversion rate on the amount field by any criteria, it depends only of the current role

There is an amount foreign (amountfx) I believe that returns the converted amount.

fx amount, it returns the converted amount in the original transaction currency, which is no good if you have multiple currencies it just sums all the numbers without matter of the currency and it makes a mess (already tested it)

you might probably need a formula column to convert the amount then, I think there is a conversion rate that you can use to do that, or create a related custom table for the corresponding transaction or fields to store the converted amount for each currency (UE)

currency consolidation can be a lot more complicated than multiplying the amount by an exchange rate