Merge pull request #24 from hhatto/imporve-result-with-apache-arrow

improve select result with apache arrow output type

examples/groonga_microblog_tutorial/3_search_error_with_pyarrow.py

@@ -0,0 +1,33 @@
+# coding: utf-8
+from poyonga import Groonga
+
+
+def _call_with_apachearrow(g, cmd, **kwargs):
+    print("[http with apache arrow response]")
+    ret = g.call(cmd, **kwargs)
+    print("status:", ret.status)
+    if cmd == "select":
+        print(type(ret.items))
+        if ret.items:
+            print("item:", len(ret.items))
+            for item in ret.items:
+                print(item)
+        else:
+            print(ret.body)
+        print("=*=" * 30)
+    else:
+        print(ret.body)
+
+
+g = Groonga()
+
+_call_with_apachearrow(
+    g,
+    "select",
+    table="Comment",  # NOTE: invalid table name
+    filter="last_modified<=1268802000",
+    output_columns="posted_by.name,comment,last_modified",
+    output_type="apache-arrow",
+    drilldown="hash_tags,posted_by",
+    drilldown_output_column="_id",
+)

poyonga/result.py

@@ -27,6 +27,7 @@ def __init__(self, data, output_type="json", encoding="utf-8", content_type=None
         if output_type == "tsv" or content_type == "text/tab-separated-values":
             # TODO: not implement
             csv.reader(StringIO(data), delimiter="\t")
+            raise NotImplementedError(f"not implement output_type: {output_type}")
         elif output_type == "msgpack" or content_type == "application/x-msgpack":
             if msgpack:
                 _result = msgpack.unpackb(data)
@@ -37,12 +38,15 @@ def __init__(self, data, output_type="json", encoding="utf-8", content_type=None
                 _result = json.loads(data)
             else:
                 _result = json.loads(data, encoding=encoding)
-        elif self._is_apache_arrow(content_type):
-            self._parse_apache_arrow(data)
-            return
+        elif output_type == "apache-arrow":
+            if self._is_apache_arrow(content_type):
+                self._parse_apache_arrow(data)
+                return
+            else:
+                raise Exception("groonga is not supported Apache Arrow output type")
         else:  # xml or other types...
             # TODO: not implement
-            pass
+            raise NotImplementedError(f"not implement output_type: {output_type}")
         self.status = _result[0][0]
         self.start_time = _result[0][1]
         self.elapsed = _result[0][2]