In § 8.1.8. "strftime() and strptime() Behavior", the documentation states that "%z" format code is a string of the form '±HHMM[SS[.uuuuuu]]' where uuuuuu is a 2-digit string giving the number of UTC offset microseconds.
I think it's a 6-digit string.

Added as part of 018d353c1c8c87767d2335cd884017c2ce12e045 and a fix regarding duplicate words for that part was added at bac2d5ba30339298db7d4caa9c8cd31d807cf081. 

Relevant format string at https://github.com/python/cpython/pull/2896/files#diff-25e2d173c84057d069b7890450714eddR214.

Relevant test case with 6-digit string for microsecond : https://github.com/python/cpython/pull/2896/files#diff-acc40bec51c7de832de3361db3edae52R309.

Table at https://docs.python.org/3.7/library/datetime.html#strftime-and-strptime-behavior could also be updated with microseconds being optional in the second column

%z | UTC offset in the form ±HHMM[SS] (empty string if the object is naive). | (empty), +0000, -0400, +1030

You can raise a GitHub PR for the same.

Thanks

FWIW more oddities with this paragraph could be fixed by:

* removing the first “and” from “HH is . . ., [and] MM is . . ., SS is . . . and uuuuuu is”,
* changing the condition for omitting “uuuuuu” from “a whole number of [minutes]” to “seconds”, and
* changing “the [SS parts are] omitted” to singular “SS part is”.

PR updated according to Martin Panter comments.

New changeset 92878829c31ab2fc71c60555ce87a5f6cbc876f0 by Alexander Belopolsky (Christophe Nanteuil) in branch 'master':
bpo-34158: Documentation UTC offset update (GH-8377)
https://github.com/python/cpython/commit/92878829c31ab2fc71c60555ce87a5f6cbc876f0

New changeset 0991b9bb94036e0f271d223c8db7d81980c76736 by Ned Deily (Miss Islington (bot)) in branch '3.7':
[3.7] bpo-34158: Documentation UTC offset update (GH-8377) (GH-9732)
https://github.com/python/cpython/commit/0991b9bb94036e0f271d223c8db7d81980c76736

I assume this can now be closed. Thanks, everyone!