Runmap rule clarification

[eyetry]

Looking at a map and haven't seen a runmap command formated like this before.

RUN("MAP_TO_RUN",ECHOIN(1,PACKAGE(InputFile1)) + " -OE1")

I get the package part fine but does the '-oe1' echo the output from output1 of "MAP_TO_RUN"? If not, when and why would I use it without a file address to go with it?

 

[eyetry]

Apparently it echo's the output from runmap back to the main map.

In these two maps both the main map output and the runmap output are 'sink'. If the output for the runmap is 50mb are the maps consuming 100mb together or are the maps sharing an address for the data in memory?

Not sure I phrased my question right.

 

[BocaBurger]

-OE1 returns the data from the RUN map. IF you are sending 50 meg. in and getting 50 meg out, it is two addresses (since the data is hopefully different). You might want to change SINK to FILE and !create (not create). Could save some RAM.



BocaBurger
<===========================||////////////////|0
The pen is mightier than the sword, but the sword hurts more!

 

[eyetry]

Actually, input is about 30mb, output is as much as 70mb so, exculding the resources functional maps consume ie... just input + output this map is consuming about 200mb sink or not. And yes the 70mb returned to the control map = the 70mb ouput of the runmap.

Wish I had the time to rewrite this whole map system now. The input file gets read at least 3 times (at the same time) just to get the initial output. The input file continues to be held open by two maps while its processed by a sequence of 5-7 additional maps. Every time a subsequent runmap is executed the output is help open locking up resources, plus, runmap output is always returned to the control map in memory + always sink.