[
https://issues.apache.org/jira/browse/AVRO-680?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14339421#comment-14339421
]
Sachin Goyal commented on AVRO-680:
-----------------------------------
Thanks [~rdblue], I have updated the pull request with my comments.
1. Here is a small java program to show why top-level maps cannot be analyzed
for non-stringiness.
{code}
import java.lang.reflect.ParameterizedType;
import java.util.HashMap;
import java.util.Map;
public class MapType {
public static Map<Integer, String> m = new HashMap<Integer, String>();
public static void main(String[] args) throws Exception {
m.put(34, "s");
// This returns false if we try to gather type-information of
the keys
System.out.println (m instanceof ParameterizedType);
// This returns true because type-information is available when
accessing it from Field
System.out.println
(MapType.class.getField("m").getGenericType() instanceof ParameterizedType);
}
}
{code}
The problem is that we only pass the type of the object and not the actual
object while getting the schema.
Due to this, for top-level maps, we do not have the "generic-type" which is
readily available in *Field*s.
Hence this issue.
\\
\\
2. I am not sure about this.
I don't think we can fall back to analyzing the type of map for empty maps,
because its type-information would be lost due to type-erasure.
So just by looking at the instance, I think its not possible to find out if its
a non-string map.
> Allow for non-string keys
> -------------------------
>
> Key: AVRO-680
> URL: https://issues.apache.org/jira/browse/AVRO-680
> Project: Avro
> Issue Type: Improvement
> Affects Versions: 1.7.6, 1.7.7
> Reporter: Jeremy Hanna
> Attachments: AVRO-680.patch, AVRO-680.patch, AVRO-680.patch,
> PERF_8000_cycles.zip, isMap_Call_Hierarchy.png, non_string_map_keys.zip,
> non_string_map_keys2.zip, non_string_map_keys3.zip,
> non_string_map_keys4.patch, non_string_map_keys5.patch,
> non_string_map_keys6.patch, non_string_map_keys7.patch,
> non_string_map_perf.txt, non_string_map_perf2.txt, original_perf.txt
>
>
> Based on an email thread back in April, Doug Cutting proposed a possible
> solution for having non-string keys:
> Stu Hood wrote:
> > I can understand the reasoning behind AVRO-9, but now I need to look for an
> > alternative to a 'map' that will allow me to store an association of bytes
> > keys to values.
> A map of Foo has the same binary format as an array of records, each
> with a string field and a Foo field. So an application can use an array
> schema similar to this to represent map-like structures with, e.g.,
> non-string keys.
> Perhaps we could establish standard properties that indicate that a
> given array of records should be represented in a map-like way if
> possible? E.g.,:
> {"type": "array", "isMap": true, "items": {"type":"record", ...}}
> Doug
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
