Github user jiadexin commented on a diff in the pull request:
https://github.com/apache/incubator-hawq/pull/972#discussion_r85676673
--- Diff:
pxf/pxf-jdbc/src/main/java/org/apache/hawq/pxf/plugins/jdbc/JdbcPartitionFragmenter.java
---
@@ -0,0 +1,284 @@
+package org.apache.hawq.pxf.plugins.jdbc;
+
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements. See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership. The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing,
+ * software distributed under the License is distributed on an
+ * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+ * KIND, either express or implied. See the License for the
+ * specific language governing permissions and limitations
+ * under the License.
+ */
+
+import org.apache.hawq.pxf.api.Fragmenter;
+import org.apache.hawq.pxf.api.FragmentsStats;
+import org.apache.hawq.pxf.plugins.jdbc.utils.DbProduct;
+import org.apache.hawq.pxf.plugins.jdbc.utils.ByteUtil;
+import org.apache.hawq.pxf.api.Fragment;
+import org.apache.hawq.pxf.api.utilities.InputData;
+
+import java.net.InetAddress;
+import java.text.SimpleDateFormat;
+import java.util.*;
+
+
+/**
+ * Fragmenter class for JDBC data resources.
+ *
+ * Extends the {@link Fragmenter} abstract class, with the purpose of
transforming
+ * an input data path (an JDBC Database table name and user request
parameters) into a list of regions
+ * that belong to this table.
+ * <p>
+ * <h4>The parameter Patterns </h4>
+ * There are three parameters, the format is as follows:<p>
+ * <pre>
+ *
<code>PARTITION_BY=column_name:column_type&RANGE=start_value[:end_value]&INTERVAL=interval_num[:interval_unit]</code>
+ * </pre>
+ * The <code>PARTITION_BY</code> parameter can be split by colon(':'),the
<code>column_type</code> current supported : <code>date,int,enum</code> .
+ * The Date format is 'yyyy-MM-dd'. <p>
+ * The <code>RANGE</code> parameter can be split by colon(':') ,used to
identify the starting range of each fragment.
+ * The range is left-closed, ie:<code> '>= start_value AND < end_value'
</code>.If the <code>column_type</code> is <code>int</code>,
+ * the <code>end_value</code> can be empty. If the
<code>column_type</code>is <code>enum</code>,the parameter <code>RANGE</code>
can be empty. <p>
+ * The <code>INTERVAL</code> parameter can be split by colon(':'),
indicate the interval value of one fragment.
+ * When <code>column_type</code> is <code>date</code>,this parameter must
be split by colon, and <code>interval_unit</code> can be
<code>year,month,day</code>.
+ * When <code>column_type</code> is <code>int</code>, the
<code>interval_unit</code> can be empty.
+ * When <code>column_type</code> is <code>enum</code>,the
<code>INTERVAL</code> parameter can be empty.
+ * </p>
+ * <p>
+ * The syntax examples is :<p>
+ *
<code>PARTITION_BY=createdate:date&RANGE=2008-01-01:2010-01-01&INTERVAL=1:month'</code>
<p>
+ * <code>PARTITION_BY=year:int&RANGE=2008:2010&INTERVAL=1</code> <p>
+ * <code>PARTITION_BY=grade:enum&RANGE=excellent:good:general:bad</code>
+ * </p>
+ *
+ */
+public class JdbcPartitionFragmenter extends Fragmenter {
+ String[] partition_by = null;
+ String[] range = null;
+ String[] interval = null;
+ PartitionType partitionType = null;
+ String partitionColumn = null;
+ IntervalType intervalType = null;
+ int intervalNum = 1;
+
+ enum PartitionType {
+ DATE,
+ INT,
+ ENUM;
+
+ public static PartitionType getType(String str) {
+ return valueOf(str.toUpperCase());
+ }
+ }
+
+ enum IntervalType {
+ DAY,
+ MONTH,
+ YEAR;
+
+ public static IntervalType type(String str) {
+ return valueOf(str.toUpperCase());
+ }
+ }
+
+ //The unit interval, in milliseconds, that is used to estimate the
number of slices for the date partition type
+ static Map<IntervalType, Long> intervals = new HashMap<IntervalType,
Long>();
+
+ static {
+ intervals.put(IntervalType.DAY, (long) 24 * 60 * 60 * 1000);
+ intervals.put(IntervalType.MONTH, (long) 30 * 24 * 60 * 60 *
1000);//30 day
+ intervals.put(IntervalType.YEAR, (long) 365 * 30 * 24 * 60 * 60 *
1000);//365 day
+ }
+
+ /**
+ * Constructor for JdbcPartitionFragmenter.
+ *
+ * @param inConf input data such as which Jdbc table to scan
+ * @throws JdbcFragmentException
+ */
+ public JdbcPartitionFragmenter(InputData inConf) throws
JdbcFragmentException {
+ super(inConf);
+ if(inConf.getUserProperty("PARTITION_BY") == null )
+ return;
+ partition_by = inConf.getUserProperty("PARTITION_BY").split(":");
+ partitionColumn = partition_by[0];
+ partitionType = PartitionType.getType(partition_by[1]);
+
+ range = inConf.getUserProperty("RANGE").split(":");
+
+ //parse and validate parameter-INTERVAL
+ if (inConf.getUserProperty("INTERVAL") != null) {
+ interval = inConf.getUserProperty("INTERVAL").split(":");
+ intervalNum = Integer.parseInt(interval[0]);
+ if (interval.length > 1)
+ intervalType = IntervalType.type(interval[1]);
+ }
+ if (intervalNum < 1)
+ throw new JdbcFragmentException("The parameter{INTERVAL} must
> 1, but actual is '" + intervalNum+"'");
+ }
+ /**
+ * Returns statistics for Jdbc table. Currently it's not implemented.
+ */
+ @Override
+ public FragmentsStats getFragmentsStats() throws Exception {
+ throw new UnsupportedOperationException("ANALYZE for Jdbc plugin
is not supported");
+ }
+
+
+
+ /**
+ * Returns list of fragments containing all of the
+ * Jdbc table data.
+ *
+ * @return a list of fragments
+ */
+ @Override
+ public List<Fragment> getFragments() throws Exception {
+ switch (partitionType) {
+ case DATE: {
+ SimpleDateFormat df = new SimpleDateFormat("yyyy-MM-dd");
+ Date t_start = df.parse(range[0]);
+ Date t_end = df.parse(range[1]);
+ int curr_interval = intervalNum;
+
+ Calendar frag_start = Calendar.getInstance();
+ Calendar c_end = Calendar.getInstance();
+ frag_start.setTime(t_start);
+ c_end.setTime(t_end);
+ while (frag_start.before(c_end)) {//||
frag_start.compareTo(c_end) == 0) {
+ Calendar frag_end = (Calendar) frag_start.clone();
+ switch (intervalType) {
+ case DAY:
+ frag_end.add(Calendar.DAY_OF_MONTH,
curr_interval);
+ break;
+ case MONTH:
+ frag_end.add(Calendar.MONTH, curr_interval);
+ break;
+ //case YEAR:
+ default:
+ frag_end.add(Calendar.YEAR, curr_interval);
+ break;
+ }
+ if (frag_end.after(c_end)) frag_end = (Calendar)
c_end.clone();
+
+ //make metadata of this fragment , converts the date
to a millisecond,then get bytes.
+ byte[] ms_start =
ByteUtil.getBytes(frag_start.getTimeInMillis());
+ byte[] ms_end =
ByteUtil.getBytes(frag_end.getTimeInMillis());
+ byte[] fragmentMetadata =
ByteUtil.mergeBytes(ms_start, ms_end);
+
+ byte[] userData = new byte[0];
+ Fragment fragment = new
Fragment(inputData.getDataSource(), null, fragmentMetadata, userData);
+ fragments.add(fragment);
+
+ //continue next fragment.
+ frag_start = frag_end;
+ }
+ break;
+ }
+ case INT: {
+ int i_start = Integer.parseInt(range[0]);
+ int i_end = Integer.parseInt(range[1]);
+ int curr_interval = intervalNum;
+
+ //validate : curr_interval > 0
+ int frag_start = i_start;
+ while (frag_start < i_end) {
+ int frag_end = frag_start + curr_interval;
+ if (frag_end > i_end) frag_end = i_end;
+
+ byte[] b_start = ByteUtil.getBytes(frag_start);
+ byte[] b_end = ByteUtil.getBytes(frag_end);
+ byte[] fragmentMetadata = ByteUtil.mergeBytes(b_start,
b_end);
+
+ byte[] userData = new byte[0];
+ Fragment fragment = new
Fragment(inputData.getDataSource(), null, fragmentMetadata, userData);
+ fragments.add(fragment);
+
+ //continue next fragment.
+ frag_start = frag_end;// + 1;
+ }
+ break;
+ }
+ case ENUM:
+ for (String frag : range) {
+ byte[] fragmentMetadata = frag.getBytes();
+ Fragment fragment = new
Fragment(inputData.getDataSource(), null, fragmentMetadata, new byte[0]);
+ fragments.add(fragment);
+ }
+ break;
+ }
+
+ return prepareHosts(fragments);
+ }
+
+ /**
+ * For each fragment , assigned a host address.
+ * In Jdbc Plugin, 'replicas' is the host address of the PXF engine
that is running, not the database engine.
+ * Since the other PXF host addresses can not be probed, only the host
name of the current PXF engine is returned.
--- End diff --
The master node's postgresql.conf file is configured with the following
parameters:
`pxf_isilon = true`
Can also be assigned to multiple pxf examples?
hd_work_mgr.c in the following code:
>
/*
* in case of remote storage, the segment host
is also where the PXF will be running
* so we set allocated->host accordingly,
instead of the remote storage system - datanode ip.
*/
if (pxf_isilon)
{
pfree(allocated->host);
allocated->host = pstrdup(host_ip);
}
>
---
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 infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
