From 07a31c71c819b2129f7eb47a8c2056de94143e0d Mon Sep 17 00:00:00 2001
From: Chris Nuernberger <chris@techascent.com>
Date: Thu, 12 Sep 2024 10:40:15 -0600
Subject: [PATCH] Fixes #414 - categorical maps are integers now by default

---
 src/tech/v3/dataset/categorical.clj       |  6 +++---
 test/data/local_date.json                 |  4 ++++
 test/tech/v3/dataset/categorical_test.clj | 10 ++++++++++
 3 files changed, 17 insertions(+), 3 deletions(-)
 create mode 100644 test/data/local_date.json

diff --git a/src/tech/v3/dataset/categorical.clj b/src/tech/v3/dataset/categorical.clj
index 0f9075b2..a09d1612 100644
--- a/src/tech/v3/dataset/categorical.clj
+++ b/src/tech/v3/dataset/categorical.clj
@@ -104,7 +104,7 @@ Non integers found: " (vec bad-mappings)))))
                m
                (set/unique (ds-base/column dataset colname)))
        colname
-       (or res-dtype :float64))))
+       (or res-dtype :int64))))
 
 
 
@@ -112,7 +112,7 @@ Non integers found: " (vec bad-mappings)))))
   "Apply a categorical mapping transformation fit with fit-categorical-map."
   [dataset fit-data]
   (let [colname (:src-column fit-data)
-        result-datatype (or (:result-datatype fit-data) :float64)
+        result-datatype (or (:result-datatype fit-data) :int64)
         lookup-table (:lookup-table fit-data)
         column (ds-base/column dataset colname)
         missing (ds-proto/missing column)
@@ -231,7 +231,7 @@ user> (ds-cat/dataset->categorical-maps catds)
         dataset (dissoc dataset src-column)
         n-elems (dtype/ecount column)
         op-space (casting/simple-operation-space (dtype-proto/operational-elemwise-datatype column))]
-    (merge dataset 
+    (merge dataset
            (->> one-hot-table
                 (lznc/map
                  (fn [[k v]]
diff --git a/test/data/local_date.json b/test/data/local_date.json
new file mode 100644
index 00000000..1a550e62
--- /dev/null
+++ b/test/data/local_date.json
@@ -0,0 +1,4 @@
+[
+	{"test": 1, "time-period": "2024-06-20"},
+	{"test": 2, "time-period": "2024-06-21"},
+	{"test": 3, "time-period": "2024-06-22"}]
\ No newline at end of file
diff --git a/test/tech/v3/dataset/categorical_test.clj b/test/tech/v3/dataset/categorical_test.clj
index 507f157f..851ea030 100644
--- a/test/tech/v3/dataset/categorical_test.clj
+++ b/test/tech/v3/dataset/categorical_test.clj
@@ -77,3 +77,13 @@
                                         (dtype/emap val-map :keyword col))))
                   (ds/categorical->number cf/categorical)
                   (ds/column "Survived")))))
+(deftest categorical-assignments-are-integers
+  (is (= #{0 1 2 3}
+         (->
+          (ds/->dataset {:x1 [1 2 4 5 6 5 6 7]
+                         :x2 [5 6 6 7 8 2 4 6]
+                         :y [:a :b :b :a :c :a :b :d]})
+          (ds/categorical->number [:y])
+          (get :y)
+          distinct
+          set))))