Revision 3181
Added by Chris Jones almost 18 years ago
| src/edu/ucsb/nceas/metacat/DocumentImpl.java | ||
|---|---|---|
| 1315 | 1315 |
updateNodeIndex(dbConn, pathList); |
| 1316 | 1316 |
} else if ( currentNode.getNodeType().equals("TEXT") ) {
|
| 1317 | 1317 |
|
| 1318 |
// Set the parent node's nodedata and nodedatanumerical to
|
|
| 1319 |
// that of the leaf TEXT node being processed, and then
|
|
| 1320 |
// traverse the parents starting from the parent node. The
|
|
| 1321 |
// xml_path_index table will be populated with ELEMENT paths
|
|
| 1322 |
// with TEXT nodedata (since it's modeled this way in the DOM)
|
|
| 1318 |
// A non-empty TEXT node represents the node data of its
|
|
| 1319 |
// parent ELEMENT node. Traverse the parents starting from
|
|
| 1320 |
// the TEXT node. The xml_path_index table will be populated
|
|
| 1321 |
// with ELEMENT paths with TEXT nodedata (since it's modeled
|
|
| 1322 |
// this way in the DOM) |
|
| 1323 | 1323 |
NodeRecord parentNode = |
| 1324 | 1324 |
(NodeRecord) nodeRecordMap.get(new Long(currentNode.getParentNodeId())); |
| 1325 |
if ( parentNode.getNodeType().equals("ELEMENT") &&
|
|
| 1326 |
!currentNode.getNodeData().equals("") ) {
|
|
| 1327 |
parentNode.setNodeData(currentNode.getNodeData()); |
|
| 1328 |
parentNode.setNodeDataNumerical(currentNode.getNodeDataNumerical()); |
|
| 1325 |
|
|
| 1326 |
if ( parentNode.getNodeType().equals("ELEMENT") ) {
|
|
| 1329 | 1327 |
|
| 1328 |
currentNode.setNodeType(parentNode.getNodeType()); |
|
| 1329 |
currentNode.setNodeName("");
|
|
| 1330 |
logMetacat.debug("Converted node " + currentNode.getNodeId() +
|
|
| 1331 |
" to type " + parentNode.getNodeType()); |
|
| 1332 |
|
|
| 1330 | 1333 |
traverseParents(nodeRecordMap, rootNodeId, |
| 1331 | 1334 |
currentNode.getNodeId(), |
| 1332 |
parentNode.getNodeId(),
|
|
| 1333 |
"", pathList, pathsFound);
|
|
| 1334 |
}
|
|
| 1335 |
currentNode.getNodeId(),
|
|
| 1336 |
"", pathList, pathsFound); |
|
| 1337 |
} |
|
| 1335 | 1338 |
} |
| 1336 | 1339 |
// Lastly, update the xml_path_index table |
| 1337 | 1340 |
if(!pathsFound.isEmpty()){
|
| ... | ... | |
| 1396 | 1399 |
String leafData = leafRecord.getNodeData(); |
| 1397 | 1400 |
float leafDataNumerical = leafRecord.getNodeDataNumerical(); |
| 1398 | 1401 |
|
| 1399 |
if (current.getNodeType().equals("ELEMENT") ||
|
|
| 1402 |
if ( current.getNodeType().equals("ELEMENT") ||
|
|
| 1400 | 1403 |
current.getNodeType().equals("ATTRIBUTE") ) {
|
| 1401 | 1404 |
|
| 1402 | 1405 |
// process leaf node xpaths |
| ... | ... | |
| 1405 | 1408 |
currentName = "@" + currentName; |
| 1406 | 1409 |
} |
| 1407 | 1410 |
logMetacat.debug("A: " + currentName +"\n");
|
| 1408 |
if ( !currentName.equals("") || !currentName.equals(null) ) {
|
|
| 1409 |
pathList.put(currentName, new PathIndexEntry(leafNodeId, |
|
| 1410 |
currentName, docid, doctype, parentId)); |
|
| 1411 |
if ( currentName != null ) {
|
|
| 1412 |
if ( !currentName.equals("") ) {
|
|
| 1413 |
pathList.put(currentName, new PathIndexEntry(leafNodeId, |
|
| 1414 |
currentName, docid, doctype, parentId)); |
|
| 1415 |
} |
|
| 1411 | 1416 |
} |
| 1412 | 1417 |
if( MetaCatUtil.pathsForIndexing.contains(currentName) && |
| 1413 |
!leafData.equals("") ){
|
|
| 1418 |
leafData.trim().length() != 0 ){
|
|
| 1414 | 1419 |
logMetacat.debug("paths found for indexing: " + currentName);
|
| 1415 | 1420 |
pathsFoundForIndexing.put(currentName, new PathIndexEntry( |
| 1416 | 1421 |
leafNodeId, currentName, docid, parentId, leafData, |
| ... | ... | |
| 1419 | 1424 |
} |
| 1420 | 1425 |
|
| 1421 | 1426 |
// process relative xpaths |
| 1422 |
if ( !currentName.equals("") || !currentName.equals(null) ) {
|
|
| 1427 |
if ( !currentName.equals("") ) {
|
|
| 1423 | 1428 |
currentName = "/" + currentName; |
| 1424 | 1429 |
currentName = currentName + children; |
| 1425 | 1430 |
} |
| ... | ... | |
| 1429 | 1434 |
} |
| 1430 | 1435 |
String path = current.getNodeName() + children; |
| 1431 | 1436 |
|
| 1432 |
if (!children.equals("")) {
|
|
| 1437 |
if ( !children.equals("") ) {
|
|
| 1433 | 1438 |
logMetacat.debug("B: " + path +"\n");
|
| 1434 | 1439 |
pathList.put(path, new PathIndexEntry(leafNodeId, path, docid, |
| 1435 | 1440 |
doctype, parentId)); |
| 1436 | 1441 |
if( MetaCatUtil.pathsForIndexing.contains(path) && |
| 1437 |
!leafData.equals("") ){
|
|
| 1442 |
leafData.trim().length() != 0 ){
|
|
| 1438 | 1443 |
logMetacat.debug("paths found for indexing: " + currentName);
|
| 1439 | 1444 |
pathsFoundForIndexing.put(path, new PathIndexEntry( |
| 1440 | 1445 |
leafNodeId, path, docid, parentId, leafData, |
| ... | ... | |
| 1444 | 1449 |
// process absolute xpaths |
| 1445 | 1450 |
if (id == rootNodeId) {
|
| 1446 | 1451 |
String fullPath = ""; |
| 1447 |
if ( !path.equals("") || !currentName.equals(null) ) {
|
|
| 1452 |
if ( !path.equals("") ) {
|
|
| 1448 | 1453 |
fullPath = '/' + path; |
| 1449 | 1454 |
} |
| 1450 | 1455 |
logMetacat.debug("C: " + fullPath +"\n");
|
| 1451 | 1456 |
pathList.put(fullPath, new PathIndexEntry(leafNodeId, fullPath, |
| 1452 | 1457 |
docid, doctype, parentId)); |
| 1453 | 1458 |
if( MetaCatUtil.pathsForIndexing.contains(fullPath) && |
| 1454 |
!leafData.equals("") ){
|
|
| 1459 |
leafData.trim().length() != 0 ){
|
|
| 1455 | 1460 |
logMetacat.debug("paths found for indexing: " + currentName);
|
| 1456 | 1461 |
pathsFoundForIndexing.put(fullPath, new PathIndexEntry( |
| 1457 | 1462 |
leafNodeId, fullPath, docid, parentId, leafData, |
| ... | ... | |
| 1571 | 1576 |
Iterator it = pathsFound.values().iterator(); |
| 1572 | 1577 |
while (it.hasNext()) {
|
| 1573 | 1578 |
PathIndexEntry entry = (PathIndexEntry)it.next(); |
| 1574 |
|
|
| 1575 | 1579 |
pstmt.setString(1,entry.docid); |
| 1576 | 1580 |
pstmt.setString(2, entry.path); |
| 1577 | 1581 |
pstmt.setString(3, entry.nodeData); |
Also available in: Unified diff
As a continued fix for http://bugzilla.ecoinformatics.org/show_bug.cgi?id=2469,
I've fixed the indexing implementation in both buildIndex() and
traverseParents(). Duane pointed out that the incorrect parent node ids
were being indexed in xml_path_index, causing some stylesheets to render
metadata incorrectly. I've changed buildindex() to index the correct parent id,
and have changed traverseParents() to only index paths that actually contain
node data (some TEXT nodes exist with empty data).