Member since
07-29-2015
535
Posts
141
Kudos Received
103
Solutions
My Accepted Solutions
| Title | Views | Posted |
|---|---|---|
| 8901 | 12-18-2020 01:46 PM | |
| 5898 | 12-16-2020 12:11 PM | |
| 4638 | 12-07-2020 01:47 PM | |
| 2797 | 12-07-2020 09:21 AM | |
| 1926 | 10-14-2020 11:15 AM |
06-13-2019
10:42 AM
Yeah I agree there is some inconsistency in behaviour here - the casting rules, especially around NULL, are too complex and inconsistent.
... View more
06-12-2019
10:48 AM
We have made some changes to decimal type resolution behaviour in recent versions. Most of them are controlled by the DECIMAL_V2 option and the default was changed starting in Impala 3.0. The best way forward is to report a bug in the Impala issue tracker - https://cwiki.apache.org/confluence/display/IMPALA/Contributing+to+Impala with the steps required to reproduce the symptoms you see. Then it can be tracked and triaged properly. This forum is not great for doing that.
... View more
06-11-2019
11:27 PM
1 Kudo
I see what you're saying there. The different from Hive isn't deliberate, but I believe this only happens in the very special case when all of the values in the CASE are NULL, without any cast. If any of them have a type, the query will execute. E.g. the below query runs and returns NULL. select 1 * (CASE WHEN '1' IS NULL THEN NULL ELSE cast(NULL as integer) END); Interestingly postgres behaves the same way as Impala psql (9.5.17, server 9.4.8)
Type "help" for help.
postgres=# select 1 * (CASE WHEN '1' IS NULL THEN NULL ELSE NULL END);
ERROR: operator does not exist: integer * text
LINE 1: select 1 * (CASE WHEN '1' IS NULL THEN NULL ELSE NULL END);
^
HINT: No operator matches the given name and argument type(s). You might need to add explicit type casts.
postgres=# select 1 * (NULL);
?column?
----------
(1 row)
postgres=# select 1 * (NULL);
?column?
----------
(1 row) We could probably file a bug if this is causing issues for you, there no specific reason it shouldn't work that I know of.
... View more
06-10-2019
07:49 AM
1 Kudo
We're working on the 5.16.2 release at the moment so I'd recommend waiting on that for the fix.
... View more
06-10-2019
07:46 AM
1 Kudo
Hi @honghan, We are aware of this issue, have a fix for it and are planning to include it in future maintenance releases for affected versions. - Tim
... View more
04-18-2019
10:09 AM
2 Kudos
If you are mainly accessing the table using Impala, I'd recommend Impala's compute stats for best performance of Impala. There are some subtle differences in the stats collected (whether they're partition or table-level). The engines can interoperate but Impala can generally generate better plans with the full set of stats from "COMPUTE STATS"
... View more
04-17-2019
06:00 PM
1 Kudo
In it's default configuration, metadata is cached until an "INVALIDATE METADATA" command evicts the table from the cache. Or until the catalog is restarted. In 5.16 and 6.1+ there are some non-default options that will evict metadata after a particular timeout. At some point these will become the defaults. Table stats are collected and stored in the hive metastore when you run a "compute stats" command. They are then just part of the table metadata.
... View more
04-17-2019
11:04 AM
1 Kudo
I think https://www.cloudera.com/documentation/enterprise/latest/topics/impala_admission.html#admission_yarn largely answers your question. There's no supported YARN/Impala integration - they each manage their own resources separately. LLAMA was an integration point but it was deprecated and removed from Impala several years ago.
... View more
04-16-2019
04:30 PM
3 Kudos
Impala caches all table metadata, so planning is generally faster once the table has been referenced by a previous query. You can see the "Planner Timeline" in the IMpala query profile to get a time breakdown of planning including metadata loading.
... View more
04-05-2019
08:54 AM
If you have more than a handful of users it becomes difficult to manage the large number of pools. Resource limits are also of limited use - you can limit the total consumption per user, but you can't guarantee that any group of users gets memory.
... View more