PEP: 279 Title: Enhanced Generators Version: $Revision$ Last-Modified: $Date$ Author: python@rcn.com (Raymond D. Hettinger) Status: Draft Type: Standards Track Created: 30-Jan-2002 Python-Version: 2.3 Post-History: Abstract This PEP introduces two orthogonal (not mutually exclusive) ideas for enhancing the generators introduced in Python version 2.2 [1]. The goal is to increase the convenience, utility, and power of generators. Rationale Python 2.2 introduced the concept of an iterable interface as proposed in PEP 234 [4]. The iter() factory function was provided as common calling convention and deep changes were made to use iterators as a unifying theme throughout Python. The unification came in the form of establishing a common iterable interface for mappings, sequences, and file objects. Generators, as proposed in PEP 255 [1], were introduced as a means for making it easier to create iterators, especially ones with complex internal execution or variable states. When I created new programs, generators were often the tool of choice for creating an iterator. However, when updating existing programs, I found that the tool had another use, one that improved program function as well as structure. Some programs exhibited a pattern of creating large lists and then looping over them. As data sizes increased, the programs encountered scalability limitations owing to excessive memory consumption (and malloc time) for the intermediate lists. Generators were found to be directly substitutable for the lists while eliminating the memory issues through lazy evaluation a.k.a. just in time manufacturing. Python itself encountered similar issues. As a result, xrange() and xreadlines() were introduced. And, in the case of file objects and mappings, lazy evaluation became the norm. Generators provide a tool to program memory conserving for-loops whenever complete evaluation is not desired because of memory restrictions or availability of data. The next steps in the evolution of generators are: 1. Add a new builtin function, iterindexed() which was made possible once iterators and generators became available. It provides all iterables with the same advantage that iteritems() affords to dictionaries -- a compact, readable, reliable index notation. 2. Establish a generator alternative to list comprehensions [3] that provides a simple way to convert a list comprehension into a generator whenever memory issues arise. All of the suggestions are designed to take advantage of the existing implementation and require little additional effort to incorporate. Each is backward compatible and requires no new keywords. The two generator tools go into Python 2.3 when generators become final and are not imported from __future__. BDFL Pronouncements 1. The new built-in function is ACCEPTED. There needs to be further discussion on the best name for the function. 2. Generator comprehensions are REJECTED. The rationale is that the benefits are marginal since generators can already be coded directly and the costs are high because implementation and maintenance require major efforts with the parser. Reference Implementation There is not currently a CPython implementation; however, a simulation module written in pure Python is available on SourceForge [5]. The simulation covers every feature proposed in this PEP and is meant to allow direct experimentation with the proposals. There is also a module [6] with working source code for all of the examples used in this PEP. It serves as a test suite for the simulator and it documents how each of the new features works in practice. The authors and implementers of PEP 255 [1] were contacted to provide their assessment of whether these enhancements were going to be straight-forward to implement and require only minor modification of the existing generator code. Neil felt the assertion was correct. Ka-Ping thought so also. GvR said he could believe that it was true. Tim did not have an opportunity to give an assessment. Specification for a new builtin [ACCEPTED PROPOSAL]: def iterindexed(collection): 'Generates an indexed series: (0,seqn[0]), (1,seqn[1]) ...' i = 0 it = iter(collection) while 1: yield (i, it.next()) i += 1 Note A: PEP 212 Loop Counter Iteration [2] discussed several proposals for achieving indexing. Some of the proposals only work for lists unlike the above function which works for any generator, xrange, sequence, or iterable object. Also, those proposals were presented and evaluated in the world prior to Python 2.2 which did not include generators. As a result, the non-generator version in PEP 212 had the disadvantage of consuming memory with a giant list of tuples. The generator version presented here is fast and light, works with all iterables, and allows users to abandon the sequence in mid-stream with no loss of computation effort. There are other PEPs which touch on related issues: integer iterators, integer for-loops, and one for modifying the arguments to range and xrange. The iterindexed() proposal does not preclude the other proposals and it still meets an important need even if those are adopted -- the need to count items in any iterable. The other proposals give a means of producing an index but not the corresponding value. This is especially problematic if a sequence is given which doesn't support random access such as a file object, generator, or sequence defined with __getitem__. Note B: Almost all of the PEP reviewers welcomed the function but were divided as to whether there should be any builtins. The main argument for a separate module was to slow the rate of language inflation. The main argument for a builtin was that the function is destined to be part of a core programming style, applicable to any object with an iterable interface. Just as zip() solves the problem of looping over multiple sequences, the iterindexed() function solves the loop counter problem. If only one builtin is allowed, then iterindexed() is the most important general purpose tool, solving the broadest class of problems while improving program brevity, clarity and reliability. Note C: Various alternative names have been proposed: iterindexed()-- five syllables is a mouthfull index() -- nice verb but could be confused the .index() method indexed() -- widely liked however adjectives should be avoided count() -- direct and explicit but often used in other contexts itercount() -- direct, explicit and hated by more than one person enumerate() -- a contender but doesn't mention iteration or indices iteritems() -- conflicts with key:value concept for dictionaries Note D: This function was originally proposed with optional start and stop arguments. GvR pointed out that the function call iterindexed(seqn,4,6) had an alternate, plausible interpretation as a slice that would return the fourth and fifth elements of the sequence. To avoid the ambiguity, the optional arguments were dropped eventhough it meant losing flexibity as a loop counter. That flexiblity was most important for the common case of counting from one, as in: for linenum, line in iterindexed(source): print linenum, line Comments from GvR: filter and map should die and be subsumed into list comprehensions, not grow more variants. I'd rather introduce builtins that do iterator algebra (e.g. the iterzip that I've often used as an example). I like the idea of having some way to iterate over a sequence and its index set in parallel. It's fine for this to be a builtin. I don't like the name "indexed"; adjectives do not make good function names. Maybe iterindexed()? Comments from Ka-Ping Yee: I'm also quite happy with everything you proposed ... and the extra builtins (really 'indexed' in particular) are things I have wanted for a long time. Comments from Neil Schemenauer: The new builtins sound okay. Guido may be concerned with increasing the number of builtins too much. You might be better off selling them as part of a module. If you use a module then you can add lots of useful functions (Haskell has lots of them that we could steal). Comments for Magnus Lie Hetland: I think indexed would be a useful and natural built-in function. I would certainly use it a lot. I like indexed() a lot; +1. I'm quite happy to have it make PEP 281 obsolete. Adding a separate module for iterator utilities seems like a good idea. Comments from the Community: The response to the iterindexed() proposal has been close to 100% favorable. Almost everyone loves the idea. Author response: Prior to these comments, four builtins were proposed. After the comments, xmap xfilter and xzip were withdrawn. The one that remains is vital for the language and is proposed by itself. Indexed() is trivially easy to implement and can be documented in minutes. More importantly, it is useful in everyday programming which does not otherwise involve explicit use of generators. Though withdrawn from the proposal, I still secretly covet xzip() a.k.a. iterzip() but think that it will happen on its own someday. Specification for Generator Comprehensions [REJECTED PROPOSAL]: If a list comprehension starts with a 'yield' keyword, then express the comprehension with a generator. For example: g = [yield (len(line),line) for line in file if len(line)>5] This would be implemented as if it had been written: def __temp(self): for line in file: if len(line) > 5: yield (len(line), line) g = __temp() Note A: There is some discussion about whether the enclosing brackets should be part of the syntax for generator comprehensions. On the plus side, it neatly parallels list comprehensions and would be immediately recognizable as a similar form with similar internal syntax (taking maximum advantage of what people already know). More importantly, it sets off the generator comprehension from the rest of the function so as to not suggest that the enclosing function is a generator (currently the only cue that a function is really a generator is the presence of the yield keyword). On the minus side, the brackets may falsely suggest that the whole expression returns a list. Most of the feedback received to date indicates that brackets are helpful and not misleading. Unfortunately, the one dissent is from GvR. A key advantage of the generator comprehension syntax is that it makes it trivially easy to transform existing list comprehension code to a generator by adding yield. Likewise, it can be converted back to a list by deleting yield. This makes it easy to scale-up programs from small datasets to ones large enough to warrant just in time evaluation. Note B: List comprehensions expose their looping variable and leave that variable in the enclosing scope. The code, [str(i) for i in range(8)] leaves 'i' set to 7 in the scope where the comprehension appears. This behavior is by design and reflects an intent to duplicate the result of coding a for-loop instead of a list comprehension. Further, the variable 'i' is in a defined and potentially useful state on the line immediately following the list comprehension. In contrast, generator comprehensions do not expose the looping variable to the enclosing scope. The code, [yield str(i) for i in range(8)] leaves 'i' untouched in the scope where the comprehension appears. This is also by design and reflects an intent to duplicate the result of coding a generator directly instead of a generator comprehension. Further, the variable 'i' is not in a defined state on the line immediately following the list comprehension. It does not come into existence until iteration starts (possibly never). Comments from GvR: Cute hack, but I think the use of the [] syntax strongly suggests that it would return a list, not an iterator. I also think that this is trying to turn Python into a functional language, where most algorithms use lazy infinite sequences, and I just don't think that's where its future lies. I don't think it's worth the trouble. I expect it will take a lot of work to hack it into the code generator: it has to create a separate code object in order to be a generator. List comprehensions are inlined, so I expect that the generator comprehension code generator can't share much with the list comprehension code generator. And this for something that's not that common and easily done by writing a 2-line helper function. IOW the ROI isn't high enough. Comments from Ka-Ping Yee: I am very happy with the things you have proposed in this PEP. I feel quite positive about generator comprehensions and have no reservations. So a +1 on that. Comments from Neil Schemenauer: I'm -0 on the generator list comprehensions. They don't seem to add much. You could easily use a nested generator to do the same thing. They smell like lambda. Comments for Magnus Lie Hetland: Generator comprehensions seem mildly useful, but I vote +0. Defining a separate, named generator would probably be my preference. On the other hand, I do see the advantage of "scaling up" from list comprehensions. Comments from the Community: The response to the generator comprehension proposal has been mostly favorable. There were some 0 votes from people who didn't see a real need or who were not energized by the idea. Some of the 0 votes were tempered by comments that the reviewer did not even like list comprehensions or did not have any use for generators in any form. The +1 votes outnumbered the 0 votes by about two to one. Author response: I've studied several syntactical variations and concluded that the brackets are essential for: - teachability (it's like a list comprehension) - set-off (yield applies to the comprehension not the enclosing function) - substitutability (list comprehensions can be made lazy just by adding yield) What I like best about generator comprehensions is that I can design using list comprehensions and then easily switch to a generator (by adding yield) in response to scalability requirements (when the list comprehension produces too large of an intermediate result). Had generators already been in-place when list comprehensions were accepted, the yield option might have been incorporated from the start. For certain, the mathematical style notation is explicit and readable as compared to a separate function definition with an embedded yield. References [1] PEP 255 Simple Generators http://python.sourceforge.net/peps/pep-0255.html [2] PEP 212 Loop Counter Iteration http://python.sourceforge.net/peps/pep-0212.html [3] PEP 202 List Comprehensions http://python.sourceforge.net/peps/pep-0202.html [4] PEP 234 Iterators http://python.sourceforge.net/peps/pep-0234.html [5] A pure Python simulation of every feature in this PEP is at: http://sourceforge.net/tracker/download.php?group_id=5470&atid=305470&file_id=17348&aid=513752 [6] The full, working source code for each of the examples in this PEP along with other examples and tests is at: http://sourceforge.net/tracker/download.php?group_id=5470&atid=305470&file_id=17412&aid=513756 Copyright This document has been placed in the public domain. Local Variables: mode: indented-text indent-tabs-mode: nil fill-column: 70 End: