Stamatis Zampetakis created HIVE-25629:
------------------------------------------

             Summary: Drop support of multiple qfiles in QTestUtil, output and 
result processors
                 Key: HIVE-25629
                 URL: https://issues.apache.org/jira/browse/HIVE-25629
             Project: Hive
          Issue Type: Task
          Components: Testing Infrastructure
    Affects Versions: 4.0.0
            Reporter: Stamatis Zampetakis
            Assignee: Stamatis Zampetakis


The current implementation of 
[QTestUtil|https://github.com/apache/hive/blob/afeb0f8413b1fd777611e890e53925119a5e39f1/itests/util/src/main/java/org/apache/hadoop/hive/ql/QTestUtil.java],
 
[QOutProcessor|https://github.com/apache/hive/blob/master/itests/util/src/main/java/org/apache/hadoop/hive/ql/QOutProcessor.java],
 and 
[QTestResultProcessor|https://github.com/apache/hive/blob/afeb0f8413b1fd777611e890e53925119a5e39f1/itests/util/src/main/java/org/apache/hadoop/hive/ql/QTestResultProcessor.java],
 has some methods and fields (maps) for managing multiple input files. However, 
*all* clients of this API, such as 
[CoreCliDriver|https://github.com/apache/hive/blob/afeb0f8413b1fd777611e890e53925119a5e39f1/itests/util/src/main/java/org/apache/hadoop/hive/cli/control/CoreCliDriver.java],
 use these classes by processing one file per run.

+Example+
{code:java}
public void runTest(String testName, String fname, String fpath) {
    ...
    qt.addFile(fpath);
    qt.cliInit(new File(fpath));
    ...
    try {
      qt.executeClient(fname);
    } catch (CommandProcessorException e) {
      qt.failedQuery(e.getCause(), e.getResponseCode(), fname, 
QTestUtil.DEBUG_HINT);
    }
    ...
}
{code}
Notice that {{qt.addFile}} will keep accumulating input files to memory 
(filename + content) while {{qt.executeClient}} (and other similar APIs) always 
operate on the last file added. Apart from wasting memory, the APIs for 
multiple files are harder to understand, and extend.

The goal of this JIRA is to simplify the aforementioned APIs by removing 
unused/redundant parts associated to multiple files to improve code 
readability, and reduce memory consumption.

+Historical note+
 Before HIVE-25625 the functionality of multiple input files was used by the 
{{TestCompareCliDriver}} but it was still useless for all the other clients. 
With the removal of {{TestCompareCliDriver}} in HIVE-25625 keeping multiple 
files is completely redundant.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

Reply via email to