We created multiple rowset readers to read data of one tablet,
after one rowset reader has reached EOF, it can be released to
reduce resource (typically memory) consumption.
As the same, we can release segment reader when it reach EOF.
SelectStmt use new TableName(null, tableRef.getAlias()) to expand star expression. tableRef.getAlias() is full name include database name and table name.
Using it as table name will generate wrong sql in CreateViewStmt.
This patch fix this problem and use correct database name and table name in expand star method.
`Select outfile into` currently only supports to export data with CSV format.
This patch extends the feature to supports parquet format.
Usage:
LocaFile:
```
SELECT citycode FROM table1 INTO OUTFILE "file:///root/doris/" FORMAT AS PARQUET PROPERTIES
("schema"="required,int32,siteid;", "parquet.compression"="snappy");
```
BrokerFile:
```
SELECT siteid FROM table1 INTO OUTFILE "hdfs://host/test_sql_prc_2019_02_19/" FORMAT AS PARQUET
PROPERTIES (
"broker.name" = "hdfs_broker",
"broker.hadoop.security.authentication" = "kerberos",
"broker.kerberos_principal" = "test",
"broker.kerberos_keytab_content" = "base64" ,
"schema"="required,int32,siteid;"
);
```
Field `schema` is required, which defines the schema of a parquet file.
Prefix `parquet.` is the parquet file properties, like compression, version, enable_dictionary.
Although the table lock can control the simultaneous modification of the table by different threads.
But it cannot control the drop operation of the table by other threads.
For example, when drop table and table update occur at the same time.
get table object by thread 1
drop table by thread 2 with table lock
update table object by thread 1
The above process is possible.
At this time, step 3 actually operates a table that no longer exists, which will eventually cause NullPointerException.
In fact, the modified table log after the drop table can be ignored. The reason is that it is meaningless to modify information on a table that no longer exists.
Fixed#5687
The version information of the tablet will be stored in the memory
in an adjacency graph data structure.
And as the new version is written and the old version is deleted,
the data structure will begin to have empty vertex with no edge associations(orphan vertex).
These orphan vertexs should be removed somehow.
Fix#5931
The reason is that sometime the method coordinate.exec() is not call when the job is timeout,
so that the query profile in this coordinate is not be initialized,
which will cause an NPE error in the execution of ExportExportingTask.
1. The partitions set by the admin repair command are prioritized
to ensure that the tablets of these partitions can be repaired as soon as possible.
2. Add an FE metric "query_begin" to monitor the number of queries submitted to the Doris.
For PR #5792. This patch add a new param `cache type` to distinguish sql cache and partition cache.
When update sql cache, we make assure one sql key only has one version cache.
When parsing memory parameters in `ParseUtil::parse_mem_spec`, convert the percentage to `double` instead of `int`.
The currently affected parameters include `mem_limit` and `storage_page_cache_limit`
The Repeat Node will change the data partition of fragment
when the origin data partition of fragment is HashPartition.
The Repeat Node will generate some new rows.
The distribution of these new rows is completely inconsistent with the original data distribution,
their distribution is RANDOM.
If the data distribution is not corrected,
an error will occur when the agg node determines whether to perform colocate.
Wrong data distribution will cause the agg node to think that agg can be colocated,
leading to wrong results.
For example, the following query can not be colocated although the distributed column of table is k1:
```
SELECT k1, k2, SUM( k3 )
FROM table
GROUP BY GROUPING SETS ( (k1, k2), (k1), (k2), ( ) )
```
According to the LRU priority, the `lru list` is split into `lru normal list` and `lru durable list`,
and the two lists are traversed in sequence during LRU evict, avoiding invalid cycles.
If query is memory exceed, a detail info where memory exceed is required.
However it is not necessary to return the entire query stack to the end user.
The query stack only needs to be printed in the be log.
