Fix parsing integers after floats in CSV upload (#40552)

src/metabase/upload.clj

@@ -175,7 +175,9 @@
                    number-regex
                    "\\s*" upload-parsing/currency-regex "?")))
 
-(defn- int-regex [number-separators]
+(defn- int-regex
+  "Matches numbers which do not have a decimal separator."
+  [number-separators]
   (with-parens
     (with-currency
       (case number-separators
@@ -184,7 +186,9 @@
         ", " #"\d[\d \u00A0]*"
         ".’" #"\d[\d’]*"))))
 
-(defn- float-or-int-regex [number-separators]
+(defn- float-or-int-regex
+  "Matches integral numbers, even if they have a decimal separator - e.g. 2 or 2.0"
+  [number-separators]
   (with-parens
    (with-currency
     (case number-separators
@@ -193,14 +197,16 @@
       ", " #"\d[\d \u00A0]*(\,[0.]+)?"
       ".’" #"\d[\d’]*(\.[0.]+)?"))))
 
-(defn- float-regex [number-separators]
+(defn- float-regex
+  "Matches numbers, regardless of whether they have a decimal separator - e.g. 2, 2.0, or 2.2"
+  [number-separators]
   (with-parens
     (with-currency
       (case number-separators
-        ("." ".,") #"\d[\d,]*\.\d+"
-        ",." #"\d[\d.]*\,[\d]+"
-        ", " #"\d[\d \u00A0]*\,[\d.]+"
-        ".’" #"\d[\d’]*\.[\d.]+"))))
+        ("." ".,") #"\d[\d,]*(\.\d+)?"
+        ",." #"\d[\d.]*(\,[\d]+)?"
+        ", " #"\d[\d \u00A0]*(\,[\d.]+)?"
+        ".’" #"\d[\d’]*(\.[\d.]+)?"))))
 
 (defmacro does-not-throw?
   "Returns true if the given body does not throw an exception."
@@ -246,17 +252,17 @@
 
 (mu/defn ^:private settings->type->check :- type->check-schema
   [{:keys [number-separators] :as _settings}]
-  (let [int?          (regex-matcher (int-regex number-separators))
+  (let [int-string?   (regex-matcher (int-regex number-separators))
         float-or-int? (regex-matcher (float-or-int-regex number-separators))
-        float?        (regex-matcher (float-regex number-separators))]
+        float-string? (regex-matcher (float-regex number-separators))]
     {::*boolean-int*   boolean-int-string?
      ::boolean         boolean-string?
      ::offset-datetime offset-datetime-string?
      ::date            date-string?
      ::datetime        datetime-string?
-     ::int             int?
+     ::int             int-string?
      ::*float-or-int*  float-or-int?
-     ::float           float?
+     ::float           float-string?
      ::varchar-255     varchar-255?
      ::text            (constantly true)}))
 

test/metabase/upload_test.clj

@@ -1926,6 +1926,25 @@
                    [2 1.0 1.0]]
                   (rows-for-table table)))))))))
 
+(deftest create-from-csv-int-and-non-integral-float-test
+  (testing "Creation should handle a mix of int and float values in any order"
+    (mt/test-drivers (mt/normal-drivers-with-feature :uploads)
+      (with-mysql-local-infile-on-and-off
+       (with-upload-table!
+         [table (let [table-name (mt/random-name)]
+                  (@#'upload/create-from-csv!
+                   driver/*driver*
+                   (mt/id)
+                   table-name
+                   (csv-file-with ["float-1,float-2"
+                                   "1,   1.1"
+                                   "1.1, 1"]))
+                  (sync-upload-test-table! :database (mt/db) :table-name table-name))]
+         (testing "Check the data was uploaded into the table correctly"
+           (is (= [[1 1.0 1.1]
+                   [2 1.1 1.0]]
+                  (rows-for-table table)))))))))
+
 (deftest append-from-csv-int-and-float-test
   (mt/test-drivers (mt/normal-drivers-with-feature :uploads)
     (testing "Append should handle a mix of int and float-or-int values being appended to an int column"