Matrix Maps¶
Matrix maps are used for data not ideal for processing one observation at a time. Matrix maps are represented by a two-dimensional matrix of floating-point values, which are accessible by a row and a column key. The keys are strings. The input tables for matrix maps are in the input directory of the playpen in a sub directory called matrices. The matrix input file needs to be present at the time the node requests the matrix map (See Sending and Waiting on Signals).
Here is an example input matrix map for foreign exchange rates:
currency,USD,EUR,JPY,GBP,CHF,CAD,AUD,HKD
USD,1.0,1.1229,0.0094,1.2450,1.0527,0.7366,0.6860,0.1290
EUR,0.8906,1.0,0.0083,1.1087,0.9375,0.6560,0.6109,0.1149
JPY,106.8000,119.9257,1.0,132.9660,112.4329,78.6682,73.2648,13.7805
GBP,0.8032,0.9019,0.0075,1.0,0.8456,0.5916,0.5510,0.1036
CHF,0.9499,1.0666,0.0089,1.1826,1.0,0.6997,0.6516,0.1226
CAD,1.3576,1.5244,0.0127,1.6902,1.4292,1.0,0.9313,0.1752
AUD,1.4577,1.6369,0.0137,1.8149,1.5346,1.0738,1.0,0.1881
HKD,7.7501,8.7026,0.0726,9.6489,8.1589,5.7087,5.3166,1.0
Note that the field names in this file do not have to comply with standard variable name rules.