Github user dwmclary commented on a diff in the pull request:
https://github.com/apache/spark/pull/3213#discussion_r20410937
--- Diff: sql/core/src/main/scala/org/apache/spark/sql/SchemaRDD.scala ---
@@ -131,6 +134,69 @@ class SchemaRDD(
*/
lazy val schema: StructType = queryExecution.analyzed.schema
+ /** Transforms a single Row to JSON using Jackson
+ *
+ * @param jsonFactory a JsonFactory object to construct a JsonGenerator
+ * @param rowSchema the schema object used for conversion
+ * @param row The row to convert
+ */
+ private def rowToJSON(rowSchema: StructType, jsonFactory:
JsonFactory)(row: Row): String = {
+ val writer = new StringWriter()
+ val gen = jsonFactory.createGenerator(writer)
+
+ def valWriter: (DataType, Any) => Unit = {
+ case(_, null) => //do nothing
+ case(StringType, v: String) => gen.writeString(v)
+ case(TimestampType, v: java.sql.Timestamp) =>
gen.writeString(v.toString)
--- End diff --
@yhaui
This is pretty pretty close. The Java APIs in, DateType and MapType are
in. However, a few things:
- If next logical action is saveAsTextFile, I'd presume this is mainly
for readers. In that case, I think the string output for timestamps is
preferable.
- UserDefinedType is proving a little tricky. I would think that we
would just add a case for UserDefinedType[_], but that's an unsupported
pattern.
On Sat, Nov 15, 2014 at 6:03 PM, Yin Huai <[email protected]> wrote:
> In sql/core/src/main/scala/org/apache/spark/sql/SchemaRDD.scala:
>
> > @@ -131,6 +134,69 @@ class SchemaRDD(
> > */
> > lazy val schema: StructType = queryExecution.analyzed.schema
> >
> > + /** Transforms a single Row to JSON using Jackson
> > + *
> > + * @param jsonFactory a JsonFactory object to construct a
JsonGenerator
> > + * @param rowSchema the schema object used for conversion
> > + * @param row The row to convert
> > + */
> > + private def rowToJSON(rowSchema: StructType, jsonFactory:
JsonFactory)(row: Row): String = {
> > + val writer = new StringWriter()
> > + val gen = jsonFactory.createGenerator(writer)
> > +
> > + def valWriter: (DataType, Any) => Unit = {
> > + case(_, null) => //do nothing
> > + case(StringType, v: String) => gen.writeString(v)
> > + case(TimestampType, v: java.sql.Timestamp) =>
gen.writeString(v.toString)
>
> If we use string for a timestamp value, the meaning of the time can be
> changed (e.g. the data is generated by a developer in a time zone and then
> it is read by another developer in another time zone). I feel using
> getTime is better (it is not very reader friendly though).
>
> @marmbrus <https://github.com/marmbrus> What do you think?
>
> â
> Reply to this email directly or view it on GitHub
> <https://github.com/apache/spark/pull/3213/files#r20406598>.
>
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
