Fixes #447

Author: cnuernber

CHANGELOG.md

@@ -1,4 +1,8 @@
 # Changelog
+# 7.038
+ * Upgrade to hamf 2.020.
+ * Fix for [issue 447](https://github.com/techascent/tech.ml.dataset/issues/447) - filter column by keyword.
+
 # 7.037
  * Nippy loading is about 2x faster in the case of large string tables.
  * Arrow read pathways support :text-as-strings? to mirror :strings-as-text? on the write side so you can save out uncompressed data in the fastest-to-read format.

deps.edn

@@ -1,6 +1,6 @@
 {:paths ["src" "resources" "target/classes"]
  :deps {;;org.clojure/clojure {:mvn/version "1.11.1"}
-        cnuernber/dtype-next {:mvn/version "10.125"}
+        cnuernber/dtype-next {:mvn/version "10.127"}
         techascent/tech.io {:mvn/version "4.31"
                             :exclusions [org.apache.commons/commons-compress]}
         org.apache.datasketches/datasketches-java {:mvn/version "4.2.0"}

test/tech/v3/dataset_test.clj

@@ -1767,6 +1767,14 @@
     (is (= :all (:print-index-range (meta (ds/sort-by-column sds :x)))))
     (is (= :all (:print-index-range (meta (ds/filter-column sds :x pos?)))))))
 
+
+(deftest issue-447-filter-column-by-keyword
+  (is (= [:a :a :a :a :a]
+         (-> (ds/->dataset {:a [:a :b :a :c :a :d :a :e :a :f]})
+             (ds/filter-column :a :a)
+             (ds/column :a)
+             (vec)))))
+
 (comment
   (require '[criterium.core :as crit])
   (def data (vec (repeatedly 100000 (fn [] {:a (rand-int 20) :b (rand) :c (rand)}))))