Strings with year abbreviation don't parse correctly

[enugentdt]

Hi there

I just started evaluating this library for use in a service. I noticed a small bit of trivia that I've found with other datetime parsing libraries.

Passing in 1y does not parse correctly. I can understand, since there is no 1m to mean one month, but year should be a unique character, at least to the english language?

As of 04/09/19, it outputs 2019-04-01 00:00:00. I'm not 100% on what unit it's deciding on, but maybe someone else will have some insight.

What makes this more interesting is if I pass en in a list of languages, it just returns None. As a matter of fact, en doesn't work at all.... Hm. I'll figure out what I'm doing wrong there

Thanks!

[Gallaecio]

There does seem to be something unexpected happening:

>>> from dateparser import parse
>>> parse('1y')
datetime.datetime(2019, 4, 1, 0, 0)
>>> parse('1y ago')
datetime.datetime(2019, 8, 1, 0, 0)
>>> parse('1 year ago')
datetime.datetime(2018, 4, 11, 12, 58, 13, 850187)
>>> parse('1y', languages=['en'])
>>> parse('1y ago', languages=['en'])
>>> parse('1 year ago', languages=['en'])
datetime.datetime(2018, 4, 11, 12, 58, 45, 962928)
>>> parse('1 y')
datetime.datetime(2019, 4, 1, 0, 0)
>>> parse('1 y ago')
datetime.datetime(2019, 8, 1, 0, 0)
>>> parse('1 y', languages=['en'])
>>> parse('1 y ago', languages=['en'])
>>> 

[noviluni]

I used the idea here: #560 (with the code from here: #561) and that is what I found:

---

In:

dateparser.parse('1y')

Log:

Missing: Month, Year
Interpretation: 1
Order: day
Used locale: es

Out:

2019-09-01 00:00:00

---

In:

dateparser.parse('1y ago')

Log:

Missing: Year
Interpretation: 1 - august
Order: day
Used locale: es

Out:

2019-08-01 00:00:00

---

In:

dateparser.parse('1y', languages=['en'])

Log:

Used locale: None

Out:

None

---

With this, it's easy to know what happens: y is skipped when using the eslocale (because it appears in skip section ines.yaml), and ago is the abbreviation for "Agosto" (August) in Spanish, so:

• First case: Equivalent to: dateparser.parse('1') (using es)
• Second case: Equivalent to: dateparser.parse('1 ago') (using es)
• Third case: When using the en locale it doesn't work because the y character is not skipped and can't be parsed.

[noviluni]

I forgot to mention that, of course, adding y as abbreviation for year in the English language will fix it: #562

[asadurski]

Beautiful! Could you just add tests supporting that theory?

[noviluni]

Beautiful! Could you just add tests supporting that theory?

Done! Is it enough?

[sparkingdark]

Can I work on it....

[Gallaecio]

@darkdebo There’s already a pull request for it: #562