[
https://issues.apache.org/jira/browse/DRILL-3535?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14727612#comment-14727612
]
ASF GitHub Bot commented on DRILL-3535:
---------------------------------------
Github user mehant commented on a diff in the pull request:
https://github.com/apache/drill/pull/140#discussion_r38555789
--- Diff:
exec/java-exec/src/main/java/org/apache/drill/exec/store/dfs/WorkspaceSchemaFactory.java
---
@@ -321,8 +327,101 @@ public DrillTable create(String key) {
return null;
}
+ private FormatMatcher findMatcher(FileStatus file) {
+ FormatMatcher matcher = null;
+ try {
+ for (FormatMatcher m : dropFileMatchers) {
+ if (m.isFileReadable(fs, file)) {
+ return m;
+ }
+ }
+ } catch (IOException e) {
+ logger.debug("Failed to find format matcher for file: %s", file,
e);
+ }
+ return matcher;
+ }
+
@Override
public void destroy(DrillTable value) {
}
+
+ /**
+ * Check if the table contains homogenenous files that can be read by
Drill. Eg: parquet, json csv etc.
+ * However if it contains more than one of these formats or a totally
different file format that Drill cannot
+ * understand then we will raise an exception.
+ * @param key
+ * @return
+ * @throws IOException
+ */
+ private boolean isHomogeneous(String key) throws IOException {
+ FileSelection fileSelection = FileSelection.create(fs,
config.getLocation(), key);
+
+ if (fileSelection == null) {
+ throw UserException
+ .validationError()
+ .message(String.format("Table [%s] not found", key))
+ .build(logger);
+ }
+
+ FormatMatcher matcher = null;
+ Queue<FileStatus> listOfFiles = new LinkedList<>();
+ listOfFiles.addAll(fileSelection.getFileStatusList(fs));
+
+ while (!listOfFiles.isEmpty()) {
+ FileStatus currentFile = listOfFiles.poll();
+ if (currentFile.isDirectory()) {
+ listOfFiles.addAll(fs.list(true, currentFile.getPath()));
+ } else {
+ if (matcher != null) {
+ if (!matcher.isFileReadable(fs, currentFile)) {
+ return false;
+ }
+ } else {
+ matcher = findMatcher(currentFile);
+ // Did not match any of the file patterns, exit
+ if (matcher == null) {
+ return false;
+ }
+ }
+ }
+ }
+ return true;
+ }
+
+ /**
+ * We check if the table contains homogeneous file formats that Drill
can read. Once the checks are performed
+ * we rename the file to start with an "_". After the rename we issue
a recursive delete of the directory.
+ * @param table - Path of table to be dropped
+ */
+ @Override
+ public void dropTable(String table) {
+
+ String[] pathSplit = table.split(Path.SEPARATOR);
+ String dirName = DrillFileSystem.HIDDEN_FILE_PREFIX +
pathSplit[pathSplit.length - 1];
+ int lastSlashIndex = table.lastIndexOf(Path.SEPARATOR);
+
+ if (lastSlashIndex != -1) {
+ dirName = table.substring(0, lastSlashIndex + 1) + dirName;
+ }
+
+ DrillFileSystem fs = getFS();
+ String defaultLocation = getDefaultLocation();
+ try {
+ if (!isHomogeneous(table)) {
+ throw UserException
+ .validationError()
+ .message("Table contains different file formats. \n" +
+ "Drop Table is only supported for directories that
contain homogeneous file formats consumable by Drill")
+ .build(logger);
+ }
+ fs.rename(new Path(defaultLocation, table), new
Path(defaultLocation, dirName));
--- End diff --
I will add a unique identifier in addition to the underscore prefix.
I modified the exception handling below to separate out
AccessControlException and the generic IOException to handle permission errors
and other errors respectively. In the case of a race with concurrent renames
the file system would throw a FileNotFoundException which would be caught by
the IOException and reported to the user.
> Drop table support
> ------------------
>
> Key: DRILL-3535
> URL: https://issues.apache.org/jira/browse/DRILL-3535
> Project: Apache Drill
> Issue Type: New Feature
> Reporter: Mehant Baid
> Assignee: Mehant Baid
>
> Umbrella JIRA to track support for "Drop table" feature.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
