I'm sure there's a simple way but how do I skip an iteration of a map function in a map/reduce script? Do I need to

break

? Netsuite makes it a pain to test these.

Just return an empty object. return {} Map stage will still trigger, but based on your condition it will not test any of he logic below and just go to next stage if any is there.

do I have to return an object or can I just use a stand-alone

return

statement?

Never tried it. Just test it out.

umm I think the standard practice would be to filter that result out at the get input stage?

At the input stage I'm returning a saved search of journal entries. For some reason Netsuite specifically made Mainline: True not work for journal entries... so in the map stage I'm trying to skip the iteration if the journal entry has already been parsed

you should be able to use the map reduce in the map you set the key as the JE internal id and in then in the reduce all the line data for that JE will be in the same reduce instance

Yes, my question was about how to skip the extra lines in the Map function. I'm trying to skip the duplicates there

yeah i think the problem you're going to run into is the map instances aren't aware of each other... so I'm not sure how the one map instance will know that the other map instance is already processing that JE

Definitely possible, I made a global array variable that i am pushing values into from the map function, and about to test to see if the global function is accessible in each instance of the map, but you're right I wouldn't be surprised if each iteration of the map function only has an empty array despite me defining it 'globally'

map instances can't see globals

or some partially filled out array for whatever that virtual machine is processing

you can use globals in MRs but only if you work withthem in summary and get input stage

Hmm okay, thanks for the info Anthony

you could maybe add a custom checkbox to the JE? and check it when you're done with that JE, and then have each map instance check for that checkbox?

that seems like a lot more work/disruptive than just running the search pre-emptively in the inputdata

true

it's only a few thousand lines and i'm only checking/storing the id value

gotcha... yeah then just do it up front in the get input would probably be easiest

i'm guessing it's an array of hashes or something?

oh you need column data? figured you'd just load the JE record?

well I mean how does the data have to be structured at the end of the getinputdata for netsuite to figure out how to iterate over it in the map function?

it just needs to be something it can iterate over.... so an array should be fine i think?

okay cool

maybe it needs to be an array of objects? so you'd do {id:x}

yeah that was what i was wondering

an array should be fine, but idk how you'd reference it, so I'd do an array of objects

their example shows-

function getInputData() {
return "the quick brown fox \njumps over the lazy dog.".split('\n'); }
// After the getInputData function is executed, the system creates the following
// key/value pairs: //
// key: 0, value: 'the quick brown fox'
// key: 1, value: 'jumps over the lazy dog.'

oh so it just makes the index the key

yeah i guess so

👍

Hmm I think that worked but it's erroring out somewhere else, do you know how to get these errors displayed? Make it fail on error

I'm not sure what you mean... if there's an error it did fail? if one map errors you want them ALL to error?

I think if it error's right now it essentially fails silently, I would like to actually see the erorr logged to the execution log, reading up on how to add that

wrap it a try/catch and log the error in the catch

Hmm yeah a try/catch could work

Yeah, aren't I doing that?

where do you use it

I didn't include that code sorry, but the results of that code is shown in my screenshot there, all the MAP STAGE CONTEXT lines

you havent shared enough code, but whatever block your context.write is in isnt running

The map context in its entirety is-

function map(context) {
      try {
        var searchResult = JSON.parse(context.value);
        // log.audit('map stage searchResult', searchResult)

        // Load the record and loop through all the lines to find customer/currency
        var journalRecord = record.load({ type: 'journalentry', id: searchResult })
        var currency = journalRecord.getValue('currency');

        var lineCount = journalRecord.getLineCount({sublistId: 'line'});
        for (var i = 0; i < lineCount; i++) {
          var account = journalRecord.getSublistValue({sublistId: 'line', fieldId: 'account', line: i});
          if (account != expenseAccountId) continue;

          var customerId = journalRecord.getSublistValue({sublistId: 'line', fieldId: 'entity', line: i});

          // Passing the required data for reduce stage
          context.write({ key: customerId, value: currency });
        }

      } catch (e) {
        log.error('hit map catch', e)
      }

      log.audit('MAP STAGE CONTEXT', context)
}

and the netsuite logs seemed to indicate that I was writing to the context with-

Audit	MAP STAGE CONTEXT	10/16/2023	11:43 am	-System-	{"type":"mapreduce.MapContext","isRestarted":false,"executionNo":1,"key":"5","value":"4327348"}

the likely explanation is that the account is not equal to the expenseAccountId

If that was the case how would I be getting that log line? Showing that it is logging distinct key/value pairs

continue is not the same as break

context.write({ key: customerId, value: currency });

is below the continue and

log.audit('MAP STAGE CONTEXT', context)

is later showing

{"type":"mapreduce.MapContext","isRestarted":false,"executionNo":1,"key":"5","value":"4327348"}

and regardless, its in a loop

so I can't write to context in a loop?

your code logs the map stage context regardless of the results of the loop

I'm not quite understanding, since it seems to be set (written?) to the context from that log statement

the keys and values of the map stage are not the keys and values of the reduce stage

I'm not following, my code seems to mirror exactly the example there and I don't see any additional information in any of the linked docs

does your code ever call mapContext.write

??? yes i copied and pasted it many times didn't I?

you have made no such check to see if that line is ever reached

So you're still saying that

if (account != expenseAccountId) continue;

is skipping every iteration of the line despite it being logged as part of the mapContext later in the map function?

you are using the value of the log

{
  "type": "mapreduce.MapContext",
  "isRestarted": false,
  "executionNo": 1,
  "key": "5",
  "value": "4327348"
}

to see a key and value, and you are assuming that has a relation to the key and value that was written

Yes, since I don't know how else that key/value would be set to context other than my context.write call?

Just return. Unless you write to map context a key and response then it won’t send any data to reduce or summarize.

Thanks Ryan, I did have context.write but the problem was that I was trying to compare to a global variable earlier in the script which obviously wasn't being set as expected, causing all the iterations to

continue

prior to hitting my context.write

Ah, yeah, everything is unique and must be passed meaning no global variables between stages.