Merge pull request #311 from mabel-dev/BUG/#310

BUG/#310

docs/SQL Reference/Working with SQL/10 Working with Timestamps.md

@@ -79,8 +79,6 @@ quarter  | ✓          | ✓       | ✓        | ✓           |
 doy      | ✘          | ✓       | ✘        | ✘           | day of year
 year     | ✓          | ✓       | ✓        | ✓           |
 
-The following convenience extraction functions also exist, however use of `EXTRACT` is recommended.
-
 ## Implicit Casting
 
 In many situation where a timestamp is expected, if an ISO1806 formatted string is provided, Opteryx will interpret as a timestamp.

docs/SQL Reference/Working with SQL/50 Relation Constructors.md

@@ -145,15 +145,15 @@ Intervals are defined quantifying one or more periods which make up the interval
 
 Recognized interval parts for the `GENERATE_SERIES` function are:
 
-Period  | Symbol
-------- | -----:
-Years   | y
-Months  | mo
-Weeks   | w
-Days    | d
-Hours   | h
-Minutes | m
-Seconds | s
+Period  | Symbol                   | Aliases
+------- | :----------------------- | ----
+Years   | **year** / **years**     | y / yr / yrs
+Months  | **month** / **months**   | mo / mon / mons / mth / mths
+Weeks   | **week** / **weeks**     | w / wk / wks
+Days    | **day** / **days**       | d
+Hours   | **hour** / **hours**     | h / hr / hrs
+Minutes | **minute** / **minutes** | m / min / mins 
+Seconds | **second** / **seconds** | s / sec / secs
 
 Where required, periods can be combined to define more complex intervals, for example `1h30m` represents one hour and 30 minutes.
 

opteryx/engine/functions/other_functions.py

@@ -14,8 +14,6 @@
 
 from pyarrow import compute
 
-from opteryx.utils import dates
-
 
 def list_contains(array, item):
     """
@@ -55,12 +53,10 @@ def search(array, item):
         return [None]
     if array_type == str:
         # return True if the value is in the string
-        # find_substring returns -1 or an index, we need to convert this to a boolean
-        # and then to a list of lists for pyarrow
-        res = compute.find_substring(array, pattern=item, ignore_case=True)
-        res = ~(res.to_numpy() < 0)
-        return ([r] for r in res)
+        return compute.match_substring(array, pattern=item, ignore_case=True)
     if array_type == numpy.ndarray:
+        # converting to a set is faster for a handful of items which is what we're
+        # almost definitely working with here - note compute.index is about 50x slower
         return (
             [False] if record is None else [item in set(record)] for record in array
         )

opteryx/utils/dates.py

@@ -22,7 +22,7 @@
 
 TIMEDELTA_REGEX = (
     r"((?P<years>\d+)\s?(?:ys?|yrs?|years?))?\s*"
-    r"((?P<months>\d+)\s?(?:mons?|mths?|months?))?\s*"
+    r"((?P<months>\d+)\s?(?:mo|mons?|mths?|months?))?\s*"
     r"((?P<weeks>\d+)\s?(?:w|wks?|weeks?))?\s*"
     r"((?P<days>\d+)\s?(?:d|days?))?\s*"
     r"((?P<hours>\d+)\s?(?:h|hrs?|hours?))?\s*"

opteryx/version.py

@@ -16,4 +16,4 @@
 2) we can import it in setup.py for the same reason
 """
 
-__version__ = "0.2.0-beta.5"
+__version__ = "0.2.0-beta.6"

tests/sql_battery/test_battery_shape.py

@@ -275,6 +275,9 @@
         ("SELECT * FROM generate_series('2022-01-01', '2022-12-31', '1 month')", 12, 1),
         ("SELECT * FROM generate_series('2022-01-01', '2022-12-31', '1 mon')", 12, 1),
         ("SELECT * FROM generate_series('2022-01-01', '2022-12-31', '1mon')", 12, 1),
+        ("SELECT * FROM generate_series('2022-01-01', '2022-12-31', '1mo')", 12, 1),
+        ("SELECT * FROM generate_series('2022-01-01', '2022-12-31', '1mth')", 12, 1),
+        ("SELECT * FROM generate_series('2022-01-01', '2022-12-31', '1 months')", 12, 1),
         ("SELECT * FROM generate_series('2022-01-01', '2022-12-31', '1 day')", 365, 1),
         ("SELECT * FROM generate_series('2020-01-01', '2020-12-31', '1day')", 366, 1),
         ("SELECT * FROM generate_series('2022-01-01', '2022-12-31', '7 days')", 53, 1),