Re: Retrieve an array using non-subscript data

> Do you mean that creating indices like you put it (with an 
> 'X' suffix at the end of your globals names) is a convention 
> when using M[umps] ?
....
> 
> Is this *really* the way to go ? Is this how it works with M[umps] ??

Globals are only implicitly indexed by their subscripts (there can be
multiple levels, though in your case there is only one).

If you want to lookup/traverse your records using some other key it's
efficient to have additional indices. M[UMPS] doesn't maintain them for
you, because it has no knowledge of the semantics of the data you're
storing at a global node. It doesn't know that piece 2 is significant.
It doesn't even know that there *are* pieces; it's just your convention
that ` is a separator of pieces of data.

An index helps you find relevant records fast. Instead of searching your
zillion records for those with a certain value of piece 2, index piece-2
values whenever you file them. It doesn't have to be a big overhead to
do this.

Of course, if you already have a zillion reecords and then decide you'd
like an index on piece 4, you have to do two things:

1. Extend your filing and deleting code so index #4 is maintained.

2. Walk the zillion records, creating an index #4 entry for each one.

Step 2 might take a while, but you only have to do it once.

As Bhaskar says, M[UMPS] is not a DBMS in its own right.

John