Yes you are. Or at least from two different angles.
System Monitor looks kind of "top-down" on the memory allocator structure and reports the usage by allocator.
M_CS_TABLES looks "bottom-up" and reports what the table data structures report.
Also M_CS_TABLES is only covering the column store tables.
- Lars
Hi All ,
While The test is getting failed for both existing connection and the new SSO relational Hana connection
java.util.ArrayList can not be cast to java.lang.String Error
Regards,
Abhi
Hello Artem,
i want to know that can't we use" ceate type tab1() as table" command inside the procedure or function. However it raises a question of what would be the data type of the output variable aswell. i tried using it inside a procedure and it raised an error that i can't use the create type inside the procedure but just want to make sure i didn't do any mistake.
thanks for sharing knowledge !
pawan
Hi Pawan,
Yes, it seems you can't call directly create type from within the SP.
But good news is that you can cheat it by executing dynamic SQL, see example below:
create procedure tempTest
as
begin
exec 'CREATE TYPE tt_test AS TABLE (col int)';
end;
call tempTest();
create procedure tempTest3 (para tt_test)
as
begin
select * from :para;
end;
create procedure tempTest2 ()
as
begin
d = select 1 col from dummy;
call tempTest3(:d);
end;
call tempTest2();
drop type tt_test;
drop procedure tempTest;
drop procedure tempTest2;
drop procedure tempTest3;
Hope it helps,
Artem
I'm currently having an issue when starting an B1 HANA DB: The index service doesn't start and the following error appears in the trace file:
Reason:
exception 1: no.1000000 (DataAccess/impl/PersistenceManagerImpl.cpp:4302)
unexpected high number of rolling back DTX found 126982 ! (free stack area of 1014952 bytes is too small); $condition$=maxTIDs <= maxIndoubtElements
I've checked for available space in the log and data volumes: Log volume is ad 93% and data volume at 7%
Has anyone had this issue?
The trace file is attached
The error message refers to the SAP HANA internal resource/memory management, not to the data or log volumes.
Apparently, there are (too) many open transactions in the system and the memory structure to keep track of them doesn't fit into the available memory.
Since you write that this happens when you try to start the instance it seems that when the instance was brought offline the last time, this was an unclean shutdown with many open transactions.
My guess here would be that it should be possible to recover this instance by providing enough stack space. But this is really a situation where I'd advise to get in touch with SAP support.
Hi Lars
Im trying to define a partitioning scheme over a TIMESTAMP column (part of the PK) that will divide the data in a partition per day fashion
I tried the following and it fails
ALTER TABLE REQUEST_INSTANCES PARTITION BY RANGE (MOD(dayofyear (START_TIME),31))
(
PARTITION 1 <= VALUES < 2,
PARTITION 2 <= VALUES < 3,
PARTITION 3 <= VALUES < 4,
PARTITION 4 <= VALUES < 5,
PARTITION 5 <= VALUES < 6,
... till day 31
)
I've checked the function syntax in the following select and it works
"select MOD(dayofyear (START_TIME),31) from REQUEST_INSTANCES;"
So why wont it let me do it in the TABLE definition ?
Any way to achieve this kind of scheme ?
hi experts,
above is the error i get if i run crystal report for SAP B1 Analytics powered by hana.
any help will be apprecited
regards
justice
It's always the sap note that one just hasn't read yet 
.
Thanks for sharing the solution!
Hi all,
can you please see the give example for CE_CONVERSION
T1=CE_COLUMN_TABLE("HANAUSER"."CE1IDEA",["FRWAE","VV010","WADAT"]);
var_out=CE_CONVERSION(:T1, [family = 'currency',
method = 'ERP', client = '800',
conversion_type = 'M', target_unit = 'EUR',
source_unit_column = " FRWAE",
reference_date = '1900-01-01',
SCHEMA="HANAUSER",
output_unit_column = "TRGCUR"]);
when doing this i am facing issues AP DBTech JDBC: [2048]: column store error: search table error: [2620] executor: plan operation failed;currency/unit conversion error,Conversion( "VV010" ), row 0: could not determine valid conversion rate for value(s) 642.6 from unit 'USD' to unit 'INR' on date 1900-01-01 00:00:00.0000
Well, this is because the partitioning function is not as versatile as general SQL functions.
The documentation lists what is possible to do here.
However, with SPS 8 the partitioning function does support date functions.
Looking at your example I wonder what additional value you expect from the MOD() function here.
The days_of_month will be re-distributed from 28-30 days to 31 days. Doesn't strike me as a big benefit.
Currently there is no dynamic PIVOT/UNPIVOT operator available in SAP HANA SQL.
I would, however, vote for it to be included into the product.
Right now, there is no real dynamic way to do this in SAP HANA.
Hi,
Thanks a lot it is solved.
BR
Sumeet
Hello Gurus,
this is my first post, so I'll have to apologize in advance for any mistakes.
We are trying to add OPInt to our HANA System. According to the guide this is done via HLM.
HLM does start but trying to 'Apply Support Package Stack' shows an error:
SAP HANA Lifacycle Manager Error
Error Message: Execution of 'Detect system landscape.' failed.
Detailed Message: System landscape description not found.
Landscape description not found in any of the expected locations [/usr/sap/HDB/lm_structure/landscapeDescription.xml, /hana/shared/HDB/lm_structure/landscapeDescription.xml]. Check SAP Note 1820588.
The same error occurs trying to view 'Installed Products and Software Components'.
However updating HLM itself from 1.0.7.12 to 1.0.7.13 is possible via HLM.
On our server there is no folder 'lm_structure'. Neither in /usr/sap/<SID>/ nor in /hana/shared/<SID>/.
So there obviously is no landscapeDescription.xml.
Note 1820588 is not helpful.
Reinstalling HLM as described in Note '1986259 - How to recreate an inconsistent landscapeDescription.xml for HANA Lifecycle Manager' does not solve the problem. An installation error occurs. However HLM seems to be installed since it starts as before.
The lifecycle management function 'Check for SAP HANA components updates' in HANA Studio is also not working as expected.
Database Version is 1.00.80.00.391861
Any helpful ideas are welcome. Thanks
Claas
There is no Dynamicway to do PIVOT.
you can use Matrix or Projection technique which are already explained in one of the blog.
Hi all
We have set up a 'Parent-Child' hierarchy in an Attribute view. The physical table in question is a simple two column table with "Parent" and "Child".
The hierarchy output worked without any problems when the record was not too high (~50,000).
But when we loaded data in the range of 1.5 billion records, there is no response coming back from HANA.
The "Select" operation on the hierarchy waits on another jobworker thread named (PopHierarchyCreate), which seems to run forever.
We are currently on SP08 and the table is not partitioned. We were not sure if partitioning would negatively affect Hierarchy performance.
Please let me know if anyone has experience with this kind of an issue and if there is any upper limit which can be processed through a 'Parent-Child' Hierarchy.
Lars Breddemann - Your expertise is very much required here
Regards
Ajay
The PopHierarchyCreate operation performs an initial traversal of the hierarchy data.
This process - afaik - does scale with the input data size (in terms of hierarchy nodes).
If the process never comes to an end, then my guess would be that there is something fishy about the data.
You may want to have support to take a more detailed look into this.
- Lars
Thanks Lars for your response.
We have sampled small chunks of the same table and the hierarchy seems to work fine without errors.
Would there be some index server settings which should be tweaked to improve hierarchy performance or would partitioning help.
We are planning to run a stored proc overnight to dump the hierarchy output into a table and check the actual response time (if it completes).
Regards
Ajay
Hi,
I am creating a filter in the calculation view based on an input parameter. User give the value of the input parameter as the begin date and I filter the data of the last 12 month from the begin date. The filter looks like follows: date <= $$begin_date$$ and data > $$begin_date$$ - 100 (date is integer in form 201509)
However I found the filter doesn't work if I put any mathematical operator inside (e.g. the minus - ). I tried created a calculated column end_date = $$begin_date$$ - 100 at first and set the filter as date <= $$begin_date$$ and data > end_date. It works but the performance is very low, much slower than I set filter as e.g. date <= $$begin_date$$ and data > 201410.
I cannot find the correct syntax to put the mathematical operator in the filter expression, do you know any secret rule to write such expression?
I considered a second parameter for the end_date or wrapper the calculation view again in sqlscript calculation view to implement the filter. But is there any better way to do that?
Thanks and regards,
Hai