std.algorithm.iFilter

[dlangBugzillaToGithub]

bearophile_hugs reported this on 2013-03-30T14:06:44Z

Transfered from https://issues.dlang.org/show_bug.cgi?id=9841

CC List

• greensunny12

Description

This is a little "challenge" on streams, LINQ, etc:

Challenge 2: Indexed Filtering

Find all the names in the array "names" where the length of the name is less than or equal to the index of the element + 1.

With LINQ:

string[] names = { "Sam", "Pamela", "Dave", "Pascal", "Erik" };
var nameList = names.Where((c, index) => c.Length <= index + 1).ToList();


In D:

    auto names2 = ["Sam","Pamela", "Dave", "Pascal", "Erik"];
    auto nameRange = iota(size_t.max)
                     .zip(names2)
                     .filter!q{ a[1].length <= a[0] }
                     .map!q{ a[1] };
    nameRange.writeln;


Using enumerate() I have suggested in Issue 5550 the D code improves:

    auto nameRange2 = names2
                      .enumerate
                      .filter!q{ a[1].length <= a[0] }
                      .map!q{ a[1] };
    nameRange2.writeln;


But in this case an iFilter is better than using enumerate+filter+map. In iFilter the filtering function is supplied by an index+item 2-tuple, so the D code becomes:

    auto nameRange2 = names2.iFilter!q{ a.length <= i };


So I suggest to add a iFilter or ifilter to Phobos. An alternative name is filterWithIndex, but iFilter is better because it reminds us that the index is the fist items of the 2-tuple.

[dlangBugzillaToGithub]

greensunny12 commented on 2018-02-10T23:27:47Z

So summarizing the status quo in D is:

---
auto names2 = ["Sam", "Pamela", "Dave", "Pascal", "Erik"];
auto nameRange2 = names2.enumerate.filter!(a => a.value.length <= a.index).map!(a => a.value);
nameRange2.writeln;
---

https://run.dlang.io/is/SzLRES

It's not too bad and I doubt that filterIndex would be accepted.