[Python-Dev] from tuples to immutable dicts (original) (raw)

Martin v. Loewis martin@v.loewis.de
23 Nov 2002 12:19:05 +0100

Armin Rigo <arigo@tunes.org> writes:

Here is a related suggestion: enhancing or subclassing tuples to allow items to be named. Example syntax:

(1,2,c=3,d=4) would build the 4-tuple (1,2,3,4) with names on the last two items.

While I agree on the need for a struct-like type, I'm strongly opposed to adding new syntax, or even make it a builtin type (i.e. one that is referenced from builtins).

Provide such a type using the standard language mechanisms.

I believe that it fills a hole between very small structures (e.g. (x,y) coordinates for points) and large structures naturally implemented as classes: it allows small structures to grow a bit without turning into an obscure 5- or 10-tuple. As a typical example, consider the os.stat() trick: it currently returns an object that behaves like a 10-tuple, but whose items can also be read via attributes for the sake of clarity.

I think the structseq type is a good starting point.

Fred once had a plan to expose structseqs to Python, to allow the creation of new structs in Python. I was suggesting that there should be a method new.struct_seq, which is called as

struct_seq(name, doc, n_in_sequence, (fields))

where fields is a list of (name,doc) tuples. The resulting thing would be similar to os.stat_result: you need to call it with the mandatory fields in sequence, and can call it with the optional fields by keyword argument.

* use 'tuple' or a new subtype 'namedtuple' or 'structure'?

No new builtins at all, please.

* the suggested syntax emphasis the use of strings for the keys, but the constructor could be more general, taking arbitrary hashable values:

YAGNI.

* how do we read the key names?

You can currently get them from the type's dict, although this contains more information than you want.

However, what do you need the key names for?

* shoud name collisions be allowed inside a namedtuple?

You mean, two fields with the same name? No.

* what about * and ** call syntaxes?

Irrelevant, since syntax extensions are not acceptable.

* dissymetries between namedtuples and dicts: operations like 'in' and iteration operates on values in tuples, but on keys in dicts...

A struct really is a sequence, not a mapping.

Regards, Martin