Fixes #401 - min/max values are incorrectly filled out.

techascent · Mar 6, 2024 · 5dfaa12 · 5dfaa12
1 parent 087214d
commit 5dfaa12
Show file tree

Hide file tree

Showing 3 changed files with 6 additions and 0 deletions.
diff --git a/src/tech/v3/libs/parquet.clj b/src/tech/v3/libs/parquet.clj
@@ -362,6 +362,8 @@ org.xerial.snappy/snappy-java {:mvn/version \"1.1.8.4\"}
                     (let [col-min (long col-min)
                           col-max (long col-max)]
                       (cond
+                        ;;Default min/max values are 0,0
+                        (== col-min col-max) :int32
                         (byte-range? col-min col-max) :int8
                         (ubyte-range? col-min col-max) :uint8
                         (short-range? col-min col-max) :int16

diff --git a/test/data/2024-03-03.parquet b/test/data/2024-03-03.parquet
diff --git a/test/tech/v3/libs/parquet_test.clj b/test/tech/v3/libs/parquet_test.clj
@@ -119,3 +119,7 @@
   (let [table (ds/->dataset "test/data/decimaltable.parquet")
         decimals (table "decimals")]
     (is (dfn/equals [3.420 1.246] decimals))))
+
+
+(deftest issue-401-paruet-missing-column
+  (is (= 4 (ds/column-count (ds/->dataset "test/data/2024-03-03.parquet")))))