Handle more date/datetime/time formats

Author: Winand

doc/source/whatsnew/v0.21.0.txt

@@ -27,6 +27,7 @@ New features
 Other Enhancements
 ^^^^^^^^^^^^^^^^^^
 
+`read_sas` function now recognizes most of frequently used date (datetime) formats in SAS7BDAT files.
 
 
 .. _whatsnew_0210.api_breaking:

pandas/io/sas/sas7bdat.py

@@ -44,8 +44,8 @@ class SAS7BDATReader(BaseIterator):
     index : column identifier, defaults to None
         Column to use as index.
     convert_dates : boolean, defaults to True
-        Attempt to convert dates to Pandas datetime values.  Note all
-        SAS date formats are supported.
+        Attempt to convert dates to Pandas datetime values.  Note that
+        some rarely used SAS date formats may be unsupported.
     blank_missing : boolean, defaults to True
         Convert empty strings to missing values (SAS uses blanks to
         indicate missing character variables).
@@ -655,9 +655,15 @@ def _chunk_to_dataframe(self):
                 rslt[name] = self._byte_chunk[jb, :].view(
                     dtype=self.byte_order + 'd')
                 rslt[name] = np.asarray(rslt[name], dtype=np.float64)
-                if self.convert_dates and (self.column_formats[j] == "MMDDYY"):
-                    epoch = pd.datetime(1960, 1, 1)
-                    rslt[name] = epoch + pd.to_timedelta(rslt[name], unit='d')
+                if self.convert_dates:
+                    unit = None
+                    if self.column_formats[j] in const.sas_date_formats:
+                        unit = 'd'
+                    elif self.column_formats[j] in const.sas_datetime_formats:
+                        unit = 's'
+                    if unit:
+                        rslt[name] = pd.to_datetime(rslt[name], unit=unit,
+                                                    origin="1960-01-01")
                 jb += 1
             elif self.column_types[j] == b's':
                 rslt[name] = self._string_chunk[js, :]

pandas/io/sas/sas_constants.py

@@ -145,3 +145,27 @@ class index:
     b"\xFF\xFF\xFF\xFE": index.columnListIndex,
     b"\xFE\xFF\xFF\xFF\xFF\xFF\xFF\xFF": index.columnListIndex,
     b"\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFE": index.columnListIndex}
+
+
+# List of frequently used SAS date and datetime formats
+# http://support.sas.com/documentation/cdl/en/etsug/60372/HTML/default/viewer.htm#etsug_intervals_sect009.htm
+# https://github.com/epam/parso/blob/master/src/main/java/com/epam/parso/impl/SasFileConstants.java
+sas_date_formats = ("DATE", "DAY", "DDMMYY", "DOWNAME", "JULDAY", "JULIAN",
+                    "MMDDYY", "MMYY", "MMYYC", "MMYYD", "MMYYP", "MMYYS",
+                    "MMYYN", "MONNAME", "MONTH", "MONYY", "QTR", "QTRR",
+                    "NENGO", "WEEKDATE", "WEEKDATX", "WEEKDAY", "WEEKV",
+                    "WORDDATE", "WORDDATX", "YEAR", "YYMM", "YYMMC", "YYMMD",
+                    "YYMMP", "YYMMS", "YYMMN", "YYMON", "YYMMDD", "YYQ",
+                    "YYQC", "YYQD", "YYQP", "YYQS", "YYQN", "YYQR", "YYQRC",
+                    "YYQRD", "YYQRP", "YYQRS", "YYQRN",
+                    "YYMMDDP", "YYMMDDC", "E8601DA", "YYMMDDN", "MMDDYYC",
+                    "MMDDYYS", "MMDDYYD", "YYMMDDS", "B8601DA", "DDMMYYN",
+                    "YYMMDDD", "DDMMYYB", "DDMMYYP", "MMDDYYP", "YYMMDDB",
+                    "MMDDYYN", "DDMMYYC", "DDMMYYD", "DDMMYYS",
+                    "MINGUO")
+
+sas_datetime_formats = ("DATETIME", "DTWKDATX",
+                        "B8601DN", "B8601DT", "B8601DX", "B8601DZ", "B8601LX",
+                        "E8601DN", "E8601DT", "E8601DX", "E8601DZ", "E8601LX",
+                        "DATEAMPM", "DTDATE", "DTMONYY", "DTMONYY", "DTWKDATX",
+                        "DTYEAR", "TOD", "MDYAMPM")

pandas/tests/io/sas/data/datetime.csv

@@ -0,0 +1,5 @@
+Date1,Date2,DateTime,DateTimeHi,Taiw
+1677-09-22,1677-09-22,1677-09-21 00:12:44,1677-09-21 00:12:43.145226,1912-01-01
+1960-01-01,1960-01-01,1960-01-01 00:00:00,1960-01-01 00:00:00.000000,1960-01-01
+2016-02-29,2016-02-29,2016-02-29 23:59:59,2016-02-29 23:59:59.123456,2016-02-29
+2262-04-11,2262-04-11,2262-04-11 23:47:16,2262-04-11 23:47:16.854774,2262-04-11