[
https://issues.apache.org/jira/browse/ATLAS-409?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15067841#comment-15067841
]
Hemanth Yamijala commented on ATLAS-409:
----------------------------------------
Hi [~dossett], tried to write a test that would simulate this condition. The
testcase was like this:
{code}
@Test
public void testCreateTableWithNoColumns() throws Exception {
String dbName = createDatabase();
String tableName = tableName();
ClassLoader classLoader = getClass().getClassLoader();
String absolutePath = new
File(classLoader.getResource("test.asvc").getFile()).getAbsolutePath();
runCommand(String.format("create table %s.%s row format serde
\'org.apache.hadoop.hive.serde2.avro.AvroSerDe\' " +
"tblproperties('avro.schema.url'='%s')", dbName, tableName,
absolutePath));
assertTableIsRegistered(dbName, tableName);
}
{code}
However, I was not able to create a test with no columns in the Hive Storage
Description. I tried to create an Avro file of this nature:
{code}
{
"namespace": "com.dsinpractice.samples.hadoop.avro",
"type": "record",
"name": "User",
"fields": []
}
{code}
But running a hive command, even in Hive, with this kind of schema is failing:
{code}
FAILED: Execution Error, return code 1 from
org.apache.hadoop.hive.ql.exec.DDLTask.
org.apache.hadoop.hive.ql.metadata.HiveException: at least one column must be
specified for the table
{code}
Can you please confirm if it is possible to define an Avro schema with no
fields? If you can upload such a file (like your test_serializer), I can try
and run the test for you. Hope this helps.
> Atlas will not import hive tables with no columns
> -------------------------------------------------
>
> Key: ATLAS-409
> URL: https://issues.apache.org/jira/browse/ATLAS-409
> Project: Atlas
> Issue Type: Bug
> Affects Versions: 0.6-incubating
> Reporter: Aaron Dossett
> Assignee: Aaron Dossett
> Attachments: ATLAS-409.patch
>
>
> Atlas won't import a Hive table with no columns (see below for an example of
> a valid hive table with no explicit columns). This is because the Atlas Hive
> Storage Descriptor class REQUIRES columns, but the Hive Table class allows
> them to be OPTIONAL.
> {code}
> CREATE TABLE example
> ROW FORMAT SERDE
> 'org.apache.hadoop.hive.serde2.avro.AvroSerDe'
> STORED AS INPUTFORMAT
> 'org.apache.hadoop.hive.ql.io.avro.AvroContainerInputFormat'
> OUTPUTFORMAT
> 'org.apache.hadoop.hive.ql.io.avro.AvroContainerOutputFormat'
> TBLPROPERTIES (
> 'avro.schema.url'='file:///path/to/the/schema/test_serializer.avsc');
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
