CSV uploads: support more date formats (#37078)

src/metabase/upload.clj

@@ -166,10 +166,10 @@
        false)))
 
 (defn- date-string? [s]
-  (does-not-throw? (t/local-date s)))
+  (does-not-throw? (upload-parsing/parse-local-date s)))
 
 (defn- datetime-string? [s]
-  (does-not-throw? (upload-parsing/parse-datetime s)))
+  (does-not-throw? (upload-parsing/parse-local-datetime s)))
 
 (defn- offset-datetime-string? [s]
   (does-not-throw? (upload-parsing/parse-offset-datetime s)))

src/metabase/upload/parsing.clj

@@ -5,6 +5,8 @@
    [metabase.public-settings :as public-settings]
    [metabase.util.i18n :refer [tru]])
   (:import
+   (java.time LocalDate)
+   (java.time.format DateTimeFormatter DateTimeFormatterBuilder ResolverStyle)
    (java.text NumberFormat)
    (java.util Locale)))
 
@@ -29,19 +31,56 @@
     :else                                (throw (IllegalArgumentException.
                                                  (tru "''{0}'' is not a recognizable boolean" s)))))
 
-(defn- parse-date
-  "Parses a date.
+(def local-date-patterns
+  "patterns used to generate the local date formatter. Excludes ISO_LOCAL_DATE (uuuu-MM-dd) because there's
+  already a built-in DateTimeFormatter for that: [[DateTimeFormatter/ISO_LOCAL_DATE]]"
+  ;; uuuu is like yyyy but is required for strict parsing and also supports negative years for BC dates
+  ;; see https://stackoverflow.com/questions/41103603/issue-with-datetimeparseexception-when-using-strict-resolver-style
+  ;; uuuu is faster than using yyyy and setting a default era
+  ["MMM dd uuuu"        ; Jan 30 2000
+   "MMM dd, uuuu"       ; Jan 30, 2000
+   "dd MMM uuuu"        ; 30 Jan 2000
+   "dd MMM, uuuu"       ; 30 Jan, 2000
+   "MMMM d uuuu"        ; January 30 2000
+   "MMMM d, uuuu"       ; January 30, 2000
+   "d MMMM uuuu"        ; 30 January 2000
+   "d MMMM, uuuu"       ; 30 January, 2000
+   "EEEE, MMMM d uuuu"  ; Sunday, January 30 2000
+   "EEEE, MMMM d, uuuu" ; Sunday, January 30, 2000
+   ])
+
+(def local-date-formatter
+  "DateTimeFormatter that runs through a set of patterns to parse a variety of local date formats."
+  (let [builder (-> (DateTimeFormatterBuilder.)
+                    (.parseCaseInsensitive))]
+    (doseq [pattern local-date-patterns]
+      (.appendOptional builder (DateTimeFormatter/ofPattern pattern)))
+    (-> builder
+        (.appendOptional DateTimeFormatter/ISO_LOCAL_DATE)
+        (.toFormatter)
+        (.withResolverStyle ResolverStyle/STRICT))))
+
+(defn parse-local-date
+  "Parses a local date string.
 
   Supported formats:
-    - yyyy-MM-dd"
+    - yyyy-MM-dd
+    - MMM dd yyyy
+    - MMM dd, yyyy
+    - dd MMM yyyy
+    - dd MMM, yyyy
+    - MMMM d yyyy
+    - MMMM d, yyyy
+    - d MMMM yyyy
+    - d MMMM, yyyy"
   [s]
   (try
-    (t/local-date s)
+    (LocalDate/parse s local-date-formatter)
     (catch Exception _
       (throw (IllegalArgumentException.
               (tru "''{0}'' is not a recognizable date" s))))))
 
-(defn parse-datetime
+(defn parse-local-datetime
   "Parses a string representing a local datetime into a LocalDateTime.
 
   Supported formats:
@@ -55,13 +94,13 @@
   (-> s (str/replace \space \T) t/local-date-time))
 
 (defn- parse-as-datetime
-  "Parses a string `s` as a LocalDateTime. Supports all the formats for [[parse-date]] and [[parse-datetime]]."
+  "Parses a string `s` as a LocalDateTime. Supports all the formats for [[parse-local-date]] and [[parse-datetime]]."
   [s]
   (try
-    (t/local-date-time (parse-date s) (t/local-time "00:00:00"))
+    (t/local-date-time (parse-local-date s) (t/local-time "00:00:00"))
     (catch Exception _
       (try
-        (parse-datetime s)
+        (parse-local-datetime s)
         (catch Exception _
           (throw (IllegalArgumentException.
                   (tru "''{0}'' is not a recognizable datetime" s))))))))
@@ -161,7 +200,7 @@
 (defmethod upload-type->parser :metabase.upload/date
   [_ _]
   (comp
-   parse-date
+   parse-local-date
    str/trim))
 
 (defmethod upload-type->parser :metabase.upload/datetime

test/metabase/upload_test.clj

@@ -167,6 +167,20 @@
            ["My favorite number is 86"   "My favorite number is 86"   vchar-type]
            ;; Date-related
            [" 2022-01-01 "                    #t "2022-01-01"             date-type]
+           [" 2022-02-30 "                    " 2022-02-30 "              vchar-type]
+           [" -2022-01-01 "                   #t "-2022-01-01"            date-type]
+           [" Jan 30 2018"                    #t "2018-01-30"             date-type]
+           [" Jan 30 -2018"                   #t "-2018-01-30"            date-type]
+           [" Jan 30, 2018"                   #t "2018-01-30"             date-type]
+           [" Feb 30, 2018"                   " Feb 30, 2018"             vchar-type]
+           [" 30 Jan 2018"                    #t "2018-01-30"             date-type]
+           [" 30 Jan, 2018"                   #t "2018-01-30"             date-type]
+           [" January 30 2018"                #t "2018-01-30"             date-type]
+           [" January 30, 2018"               #t "2018-01-30"             date-type]
+           [" 30 January 2018"                #t "2018-01-30"             date-type]
+           [" 30 January, 2018"               #t "2018-01-30"             date-type]
+           [" Sunday, January 30 2000"        #t "2000-01-30"             date-type]
+           [" Sunday, January 30, 2000"       #t "2000-01-30"             date-type]
            [" 2022-01-01T01:00 "              #t "2022-01-01T01:00"       datetime-type]
            [" 2022-01-01t01:00 "              #t "2022-01-01T01:00"       datetime-type]
            [" 2022-01-01 01:00 "              #t "2022-01-01T01:00"       datetime-type]