Force to use zonemap for collecting string type min max.
String type is not using zonemap for min max, because zonemap value at BE side is truncated at 512 bytes which may cause the value not accurate. But it's OK for statisitcs min max, and this could also avoid scan whole table while sampling.
The first time load data to a partition, we need to analyze the partition columns even when the health rate is high. Because if not, the min max value of the column may not include the new partition values, which may cause bad plan.
Sample analyzing need to get row count by using table.getRowCount(). This method is not updated in real time, which may cause the sample task to scan whole table.
This pr is to fix this. Set the flag that indicate the analyze job is for an empty table and skip scan the table. Meanwhile, don't reset updatedRows in this case.
Set hugeTableAutoAnalyzeIntervalInMillis = 0 because all default huge table size has been set to 0.
Drop stats use IN predicate to filter the column stats to delete. The default length of IN predicate is 1024, drop table stats with more than 1024 columns may fail.
This pr is to split the delete sql based on the IN predicate length.
Before, we didn't update the jobType info in ColStatsMeta. This will case the jobType always be the type
when it first be set. For example, if we manually analyzed table, the jobType will always be MANUAL,
even if this table is auto analyzed again later.
1. not collect partition stats anymore
2. merge insert of stats
3. delete period collector since it is useless
4. remove enable_auto_sample
5. move some config related to stats to global session variable
Before this PR, when analyze a table, the insert count equals column count times 2
After this PR, insert count of analyze table would reduce to column count / insert_merge_item_count.
According to my test, when analyzing tpch lineitem, the insert sql count is 1
It has following problems before this PR
use count(*) to check if all column analyzed
return directly when fe count > 1
Co-authored-by: AKIHA <cyborgz1999@example.com>
1. Add a abstraction for column stats status which is required so furthur optimization and feature development
2. Enable analyze test in p0 that disabled unexpectedly before
It's caused by we used same query id for multiple queries of same olap analyze task, but many structures related to query execution depends on query id.
1. Analyze with sample automatically when table size is greater than huge_table_lower_bound_size_in_bytes(5G by default). User can disable this feature by fe option enable_auto_sample
2. Support grammer like `ANALYZE TABLE test WITH FULL` to force do full analyze whatever table size is
3. Fix bugs that tables stats doesn't get updated properly when stats is dropped, or only few column is analyzed
