topic Re: Slicing Associative Arrays? in Discussions

Slicing Associative Arrays?

hogi — Sat, 18 Jan 2025 05:55:34 GMT

Fox matrix indexing, one can use 0 and get a full slice:

m = [ 1 2 , 3 4];
m[0,1]; // [1, 3]

For Data tables this works as well:

test = new table("test", 
add rows(10), 
new column("col1",Character, set each value({"A","B","C"}[random integer(3)])),
new column("col2", set each value(random uniform(5))));
test[4,0]

For lists of lists, this is not possible:

list= {{"A", 3},{"B", 5}};
list[0,1]

Sure, there are many lists of lists where such slicing is not possible, like {{"A", 3},{"B"}}.

On the other hand, there ARE many lists of list that would allow it.

Where it gets really interesting: Associative arrays of associative arrays.
Like with lists, the concept of Associative Arrays is more flexible than what we can do with a matrix or data table - so slicing will not work for them. But like with lists, there ARE Associative arrays where slicing will work ...

Is there a function in JSL, something like myAssociativeArray[0, "weight"], which does this job?

Names Default to Here(1);
dt = Open( "$SAMPLE_DATA/Big Class Families.jmp" );
dt << Save( "$DOWNLOADS/test.json" );

// for sake of simplicity we accept to lose 1x ROBERT
myIndex=dt[0,"name"]

myAssociativeArray = Associative Array(myIndex,ParseJSON(Load text file ( "$DOWNLOADS/test.json")));

// cool, but not what we want
mapToWeight = Transform Each({{key, value}}, aa, value["weight"]);

SliceAssociativeArray= Function({aa, indexname},
	Transform Each({value}, aa << get values, value[indexname])
);

SliceAssociativeArray(myAssociativeArray ,"weight") //myAssociativeArray [0, "weight"]

Re: Slicing Associative Arrays?

jthi — Sat, 18 Jan 2025 06:27:26 GMT

Different data structures have different uses, for associative array (I think) it means storing unique keys and getting the values for those keys fast (it stores key-value pairs). Transform Each is one good option to use, just change the output

weights = Transform Each({{key, value}}, aa, Output("List"), 
	value["weight"]
);

Re: Slicing Associative Arrays?

hogi — Sat, 18 Jan 2025 06:35:03 GMT

Another cool trick - very useful :)

From the useability, I think

weights = Transform Each({{key, value}}, aa, Output("List"), 
	value["weight"]
);

and

	Transform Each({value}, aa << get values, value["weight"])

from the original post are on a comparable value

... quite far from aa[0, "weight"]

Just imagine that you want to get all the aa[0,"2024",0,0,"date"].

Re: Slicing Associative Arrays?

jthi — Sat, 18 Jan 2025 06:48:06 GMT

I'm not sure what this should "query" aa[0,"2024",0,0,"date"].If you only wish to get "0" indices, you could store your values to associative array in such a way, that your keys are column names and values are lists of the row values.

Names Default To Here(1); 

dt = open("$SAMPLE_DATA/Big Class.jmp");

aa = Associative Array();
For Each({colname}, dt << Get Column Names("String"),
	aa[colname] = As List(Column(dt, colname) << get values);
);

Generally if I want to store my data in table like format which can be accessed like a table, I would just use (private) JMP data table.

Re: Slicing Associative Arrays?

hogi — Sat, 18 Jan 2025 06:52:28 GMT

good point : ):)