[
https://issues.apache.org/jira/browse/PIRK-78?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15603061#comment-15603061
]
ASF GitHub Bot commented on PIRK-78:
------------------------------------
Github user tellison commented on a diff in the pull request:
https://github.com/apache/incubator-pirk/pull/112#discussion_r84772832
--- Diff: src/main/java/org/apache/pirk/schema/query/QuerySchemaLoader.java
---
@@ -226,50 +215,28 @@ public QuerySchema loadSchema(InputStream stream)
throws IOException, PIRExcepti
Element elements = (Element) elementsList.item(0);
LinkedHashSet<String> elementNames = new LinkedHashSet<>();
- int dataElementSize = 0;
NodeList nList = elements.getElementsByTagName("name");
for (int i = 0; i < nList.getLength(); i++)
{
Node nNode = nList.item(i);
if (nNode.getNodeType() == Node.ELEMENT_NODE)
{
- // Pull the name
- String queryElementName =
nNode.getFirstChild().getNodeValue().trim();
- if (!dataSchema.containsElement(queryElementName))
- {
- throw new PIRException("dataSchema = " + dataSchemaName + " does
not contain requested element name = " + queryElementName);
- }
- elementNames.add(queryElementName);
- logger.info("name = " + queryElementName + " partitionerName = " +
dataSchema.getPartitionerTypeName(queryElementName));
-
- // Compute the number of bits for this element.
- DataPartitioner partitioner =
dataSchema.getPartitionerForElement(queryElementName);
- int bits =
partitioner.getBits(dataSchema.getElementType(queryElementName));
-
- // Multiply by the number of array elements allowed, if applicable.
- if (dataSchema.isArrayElement(queryElementName))
- {
- bits *=
Integer.parseInt(SystemConfiguration.getProperty("pir.numReturnArrayElements"));
- }
- dataElementSize += bits;
-
- logger.info("name = " + queryElementName + " bits = " + bits + "
dataElementSize = " + dataElementSize);
+ elementNames.add(nNode.getFirstChild().getNodeValue().trim());
}
}
+ schemaBuilder.setQueryElementNames(elementNames);
// Extract the filter, if it exists
- String filterTypeName = NO_FILTER;
if (doc.getElementsByTagName("filter").item(0) != null)
{
- filterTypeName =
doc.getElementsByTagName("filter").item(0).getTextContent().trim();
+
schemaBuilder.setFilterTypeName(doc.getElementsByTagName("filter").item(0).getTextContent().trim());
}
// Create a filter over the query elements.
- Set<String> filteredNamesSet = extractFilteredElementNames(doc);
- DataFilter filter = instantiateFilter(filterTypeName,
filteredNamesSet);
+
schemaBuilder.setFilteredElementNames(extractFilteredElementNames(doc));
// Extract the additional fields, if they exists
- HashMap<String,String> additionalFields = new HashMap<String,String>();
+ Map<String,String> additionalFields = new HashMap<String,String>();
--- End diff --
Yep, I didn't spot that one - fixed.
> Create a new QuerySchema builder to separate from representation
> ----------------------------------------------------------------
>
> Key: PIRK-78
> URL: https://issues.apache.org/jira/browse/PIRK-78
> Project: PIRK
> Issue Type: Improvement
> Reporter: Tim Ellison
>
> Presently, Pirk has a QuerySchemaLoader that reads an XML formatted stream
> and builds a new QuerySchema.
> In order to separate the representation of the schema in XML from the logic
> of assembling a new schema such as validating the values, a new
> QuerySchemaBuilder is to be introduced. The QuerySchemaLoader will then
> delegate the creation to the new builder, and other representations can also
> use the builder to create a new QuerySchema without having to go via XML.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
