Member since
10-16-2013
307
Posts
77
Kudos Received
59
Solutions
My Accepted Solutions
Title | Views | Posted |
---|---|---|
11157 | 04-17-2018 04:59 PM | |
6115 | 04-11-2018 10:07 PM | |
3523 | 03-02-2018 09:13 AM | |
22121 | 03-01-2018 09:22 AM | |
2633 | 02-27-2018 08:06 AM |
05-27-2016
12:32 AM
It is not possible to do that today, because Impala UDFs can only produce scalar values (and can only take scalar arguments). We do intend to extend the UDF support to allow complex argument and return types, but there is no concrete timeline yet.
... View more
05-26-2016
12:10 AM
You're welcome. Sorry for the inconvenience.
... View more
05-25-2016
11:40 PM
You are not doing anything wrong. I should have been more clear in my description of the workaround, sorry. It seems like what you want is CHAR comparison semantics, but Impala currently does not provide that directly. To avoid confusion, I'd recommend using the STRING type consistently. However, comparison of STRINGs are whitespace-sensitive, so you will have to work around that, e.g., by manually trimming the whitespace when doing comparisons. So in your example, you'd have to change your query to: select * from test where trim(c1) = 'ABC';
... View more
05-25-2016
10:49 AM
Thanks for your report. This is a known issue: https://issues.cloudera.org/browse/IMPALA-1652 The simplest workaround is to use the STRING data type instead.
... View more
05-25-2016
10:10 AM
1 Kudo
Mystery solved, I had a typo as well 🙂 The option should be: JAVA_TOOL_OPTIONS="-Xmx8g" (TOOL instead of TOOLS)
... View more
05-25-2016
10:06 AM
Thanks for following up with more details. We're investigation and will get back to you as soon as we have more info.
... View more
05-24-2016
10:19 PM
Setting the JAVA_TOOL_OPTIONS like you did via the CM safety valve is the preferred way and should work. Did you perhaps have a typo? JAVA_TOOLS_OPTIONS="-Xmx8GB" should be JAVA_TOOLS_OPTIONS="-Xmx8g"
... View more
05-23-2016
11:00 PM
1 Kudo
What you are looking for are table functions (or user-defined table functions) like split(). Unfortunately, Impala does not yet support table functions. You should consider using Hive which has the split() function to transform the data into a more read-friendly format. I understand that some data is produced in such a nested-delimited format, but you depending on your use case it could be wasteful to split and unnest the string data for every read query.
... View more
05-20-2016
10:36 PM
Impala does not support dynamic SQL. I am not 100% sure about Hive, but I doubt Hive support it.
... View more
05-19-2016
09:21 AM
1 Kudo
Impala's Parquet scaner materializes all columns required by the query in memory, and then evaluates predicates to discard non-matching rows. If you have many columns, then this materialization is expensive. This issue is described in https://issues.cloudera.org/browse/IMPALA-2017 and is active area of work to improve performance. Alex
... View more