Hello Experts

There are many discussions that happened in the forum about the memory utilization and limits.

I have been seeing this symptom in our HANA systems quite a number of times and posting this for clarification:

During the business hours the memory usage is used to the peak and usually settles down during non-business hours.

Below overview page shows the Peak Used Memory as 129.08 GB vs. Allocated GB is only 128.

What is the concept here of Peak Used Memory exceeding the Allocation Limit ?

But after couple of minutes this came back to 125.69 GB. And when I ran an SQL query from the snapshots go the below output and I will never get the Peak Used memory as 129.08 or 125.69 GB and it bit lesser that what it shows up in the Studio:

What is correct Peak used memory here ? Which one is true ? Where I do get the correct and reliable information ?

Thanks

Srikar

Hi All,

Had this query and have searched for the same in scn as well.

http://scn.sap.com/thread/3334993    

Now if we have a HANA box and will there be any need to partition the table if we have a single node box.

There is a limitation in HANA for the 2 billion records right, but i wasnt able to find the same from "select *from M_system_limits". Will this hold good.

How would the optimiser handle the partitioned table since its just a single node hana. Will it make sense if we have just one index server process to have the table partitioning. Please share your experiences.

We are still in the initial stage we are not quite sure as to how things will do related to performance. So wanted to seek some opinions regarding the same of partitioning of the table on a single node.

Since most of the column store will be in memory anyways reading wouldnt be a problem i guess, but delta merge might be a costly operation if the table is really huge and unpartitioned? What do you think experts?

Dears,

When trying to upgrade HANA from 9.0 PL16 to 9.1 PL05 it shows:

how we can solve the issue ?

Hi all,

I have to implement the following logic in an HANA calculation view preferably in Graphical calculation view.

I have a field like BKTXT which contains either number or text or both and datatype of the field is VARCHAR.

Requirement is to create a calculated column:

• If BKTXT starts with a number, convert the numeric part to a number (ex. If BKTXT= “345 – Absolute”, result should be 345)
• If BKTXT contains only text or it starts with a text and not a number output should be 1 (ex. If BKTXT= “SCM DV” or “SCM DVI 57”, result should be 1)
• If BKTXT is Blank or ‘Null’, then result should be considered as Zero (0)

How can I implement this logic?

Here was the algorithm I thought of though unable to implement so far:

1. To find whether the first character is between 0 to 9 or not.
2. If no, then result is 1.
3. If yes, then take that. Then check for the 2^nd one and so on till it is between 0 to 9 and concatenate the same.

I am facing challenge to identify a numeric field in character field. If I can do that also how to run a loop here to take our a 3 digit or more number from the text field

Thanks in advance.

Hi All,

We are using SP8 version of HANA and created all the tables using hdbtable concept. Now we have a requirement to give foreign key relationships between the tables. I found in many posts that foreign key relationships is currently not supported and as an alternative people are asking to create a procedure and write the code. But in developer guide I saw an option to give foreign key constraints as below. When i try to create an .xsodata object and activate it with the code, it is not generating any constraint. I performed test to check the referential integrity, But it did not work. Please let me know if I am missing any  setting or step.

Code used :

service { "sample.odata::customer" as "Customers"; "sample.odata::order" as "Orders"; association "Customer_Orders" principal "Customers"("ID") multiplicity "1" dependent "Orders"("CustomerID") multiplicity "*"; }

Note : Code wise, there is no problem for me with syntax.

Hi all,

we've just made a fresh new development Hana instance on AWS. It works, except that in the Studio administration console it says

SQL error while updating the administration overview

Alerts cannot be displayed

On the bottom it also says

Statistics server not found as service

and various other errors. In the Landscape tab there is indeed no statisticsserver service running. In the configuration nameserver.ini/statisticsserver is active = false, setting it to true does not change anything. Calling

SELECT * FROM _SYS_STATISTICS.STATISTICS_PROPERTIES where key = 'internal.installation.state';

gets me

* 3589: remote query execution failure: OpenChannelException at network layer:

  message: an error occured while opening the channel

  info:    connection refused

  param:   127.0.0.1:30005

SQLSTATE: HY000

So it looks like this instance is not really well. Can this be fixed somehow?

Cheers,

-- Micha

Hi All,

I'm facing some problem while connecting to the HANA System. It's a stand alone HANA DB.

Machine A: HANA DB (Suse Linux)

Machine B: HANA Studio (Windows Server 2008)

I have done the following trying to resolve this but no luck

1. Put an entry in host files in both the machines

2. HANA Server is up and running

3. Tried to restart and start the services, still not able to resolve

Given below the screenshot of the error message

Any idea what could be the issue here?

Regards,

Skugan V

Hello,

I have an error when using arrays in sql coding in my calculation view

My target is to have the different values of the array.

In this sample coding, I want to have the first value in the array (original coding is much more complex)

I get my data from table ZTAB, only column COL which is NVARCHAR(1)

Coding is as bellow:

BEGIN
DECLARE ma      NVARCHAR(1) ARRAY;
DECLARE ar      NVARCHAR(1) ARRAY;
DECLARE lv_ar   NVARCHAR(1);
lt_tab = select distinct col from ztab;                  
ar := ARRAY_AGG(:lt_tab.col);
lv_ar := :ar[1];                                          
var_out = UNNEST(:ma, :ar) AS ("MANDT", "COL");              
END ;

The problem come from line 7 (lv_ar := :ar[1];), if I delete line 7 activation will be ok...

For sure it's a conversion error in line 7 but I don't understand why...

Error text:

Internal deployment of object failed;Repository:

Encountered an error in repository runtime extension;Internal

Error:Deploy Calculation View: SQL: transaction rolled back by an internal error:

column store error: [34011] failed to save calculation scenario :

The following errors occured: Inconsistent calculation model (34011)

Thank you in advance for your help.

Moez.

Hi All,

I have a requirement to determine the first day of the week based on the locale.

For example,

• in US, the first day of this week is Sunday, March 29
• in Germany, the first day of this week is Monday, March 30

I know I can get the first day of the week using function WEEKDAY, so I tried below SQL which works, but since WEEKDAY is hardcoded Monday(0) to Sunday(6), the results are always giving the first day of the week as Monday.

from dummy;

Any ideas how to make this locale dependent?

Thanks and best regards,

Cosmin

Hi,

We have requirement to show Number of active customers by Calendar date in reporting. So using Valid From and Valid to we are determining number of Active Customers for the particular calendar date for last two months ( for every date we need to find respective values in view).

For this purpose we are creating Scripted calculation view. But we cannot get full result set  we are getting only last loop results. My assumption is this is happening because :table variable is reference and it contains only last result set.

Is it possible to append/insert result set to some temporary table and only then assign to var_out?

Is there other way of achiving this requirement?

Following code is added:

DECLARE lv_analisys_start date;

DECLARE lv_today date;

DECLARE lv_date date;

DECLARE lv_int integer := 0;

DECLARE rowcount integer; 

lv_analisys_start := to_date(left(to_dats(now()), 6)-1);

lv_today := current_date;

dates_table = select"DATE_SQL"from"_SYS_BI"."M_TIME_DIMENSION"

where date_sql >= :lv_analisys_start AND date_sql <= :lv_today;

SELECTcount(*) into rowcount from :dates_table;

WHILE :lv_int < rowcount DO

select"DATE_SQL"into lv_date from :dates_table LIMIT 1 OFFSET :lv_int;

table =  select"PRDNR", "CITY",

sum(casewhen to_date(left("VALFR",8)) <= lv_date

AND to_date(left("VALTO",8)) >= lv_date then 1

else 0

end) AS"CM_ACTIVE_CUSTOMERS", lv_date asDATE

FROM"_SYS_BIC"."sap.development/CA_CUSTOMER_DATA"

groupby"PRDNR", "CITY";

lv_int := :lv_int + 1;

var_out = select * from :table;

ENDWHILE;

Thanks,

Markus

Hi all,

I am developing a HANA Native application , and it is reachable through the web with a web browser.

What I want to ask is whether there is a (preferable efficient) way to make it applicable for mobile versions like Android / ios.

Thank you,

Christoforos Verras

Hi all,

I'm copying some tables from one instance to another one using EXPORT on one side and IMPORT WITH REPLACE DATA ONLY on the other side. I'm using the DATA ONLY clause because without it all permissions on the target table are deleted. Now this works fine, until I've included a table with an index, now the import says

* 2048: column store error: table import failed:  [30139] Could not import data only due to metadata mismatch;Data only import is not possible: create statement differs,object=I2S:CATSen SQLSTATE: HY000

When I remove the DATA ONLY clause or when I delete the index on both instances the import works, but I'm losing the index or the privileges. NB the tables are identical and so is the index, the only difference is that the source Hana is SPS7 and the target Hana is SPS9. Is there a way to make this work? I want to copy only data and keep both the index and the granted privileges.

Cheers,

-- Micha

Hello ,

Anyone cleared HANA Installation Certification ? Can anybody share his/her experience in preparing this examination with sample questions

Thanks and Regards,

Hemant Kumar

Would you please suggest on Migration from SAP ECC 6.0 to SAP HANA.

Having DB Size 800GB deploy HANA on premises or HANA on cloud, which one is viable as far Performance and Investment point.

Would you please provide more info on this.

All,

I am a functional consultant having experience in modules MM/PM/SD. I have no ABAP programming experience, but before becoming a SAP consultant I did program in JAVA and JavaScript quite a lot. When I read the HANA documentation, I soon realize that there could be some skillets I may posses that can be used to build some native (XSJS) and non-native (JAVA) HANA applications. This interest lead me to creating a trial account and I was able to build a vary basic application using both XSJS and Java.

Here are some questions that I was hoping to get out of this discussion ..

1. Would there be a significant performance difference between Native and Non-Native HANA applications ? Does anyone have experience comparing performances in a production environment ?
2. If I had to tinker around with HANA and build some simple applications that read and created transactions in a HANA SAP Business One environment (ERP), how would I go about it ? Are there XSJS APIs available to connect to the ERP system (Similar to a BAPI) ?
   • Example application 1  - List PM01 work orders created after a certain date, that does not have status TECO
   • Example application 2  - If a PM01 work order meets a certain criteria, trigger TECO from my HANA application
3. I did hear about the AWS developer account, but thought I will hear from someone that has done something like this.

Thanks for your help ...

Jose

Dear experts,

I am having an issue due to the time change in our SAP HANA DATABASE. We have a SAP BW connected with a SAP Hana Database, BW is on Windows, and HANA is on SuSe.

During the time change, it seems there was an issue in the SuSe server which didn't automatically change the time. So this, I believe, created inconsistencies in the database.

I restored the SAP Hana database to before the time change and after a SPS update that I performed, but I still cannot get the SAP BW application up. Everytime I do, it stops a few minutes laters with ZDATE_LARGE_DIFFERENCE errors, and saying "there's a large difference between the application server and database".

Now, the application server and the SuSe host have the same hour (WEST correctly). However, if I look at the logs on SAP HANA (for example, availability.log)... I can see that the time is wrong, it is 17:00 instead of 18:00 (for example).

What could be the cause? I mean, I don't know what is controlling the time that SAP Hana database is internally taking!

I have executed the following command and I see the timezone is set as "WEST" (which is correct):

- select * from M_HOST_INFORMATION where upper(KEY) like '%TIMEZONE%'

Any advise is welcome!!

Thank you!

I am part of SAP startup focus program and was searching for some guide to replicate the HANA model that has been created on one HANA DB to another. I require this as part of setting up similar data models in different HANA servers to cater to different clients. It would be great if I can get some assistance or any guide online for the same.

Thanks in advance

Hi All,

We are currently facing a challenge in our project around security. We are working on HANA 1.0 SP07 rev73 and using tableau for visualisation.

There is content that need to be secured and at a package and schema level is fine (we've created a separate schema for this data). The roles we've created do their job with some gaps. The users that don't have these roles can't query the tables in the secure schema and they can't see the information models created in the secure package. however, the challenge is when building information models around this data, all these views get activated under the _SYS_BIC schema and users that don't hold the role for this secure data can still execute them and see them because of their select privilege on _SYS_BIC. Tableau's connector is done by selecting the column view from the _SYS_BIC schema which makes these views visible.

I understand that this happens because of having SELECT access on _SYS_BIC for all users. As far as I know there is no star (*) functionality to segregate access per models in each package and that we'll need to give select access to each separate model to do so. This, as you can see, is not very efficient from a maintenance point of view because for every model that developers build security will have to add it to their role for them to test it.

Other alternative we've thought of was to write a stored procedure to manage this granular granting every 5 minutes which will assign the latest models created to the corresponding role.

So, has anyone out there had a similar scenario and worked out an effective solution to it?

all suggestions are welcome.

Best regards,

Christian.

Hi,

We just gone live with Business Suite on HANA last month. With few weeks of backup, the disk space for backup is not enough. We would like to extend the disk space.

May I know what is the command to extend the disk space? Do I need to stop the SAP and HANA before extend?

Will I lose the existing data in the backup folder?

Thank you.

regards,

zl

Hello everyone,

I am having trouble with using a result view from within a procedure.

The definition of the result view looks like this:

create procedure NEO_CCG502K10CQSC8NX8DNT6J7MA.GET_RULE_050_PROC_V1(

    in i_tenant_id NVARCHAR,

    out o_result NEO_CCG502K10CQSC8NX8DNT6J7MA.RULE_050_TYPE)

    reads sql data

    with result view NEO_CCG502K10CQSC8NX8DNT6J7MA.RULE_050_PROC_VIEW

    as ...

I can call this result view from the SQL console the following without any problems:

select * from NEO_CCG502K10CQSC8NX8DNT6J7MA.RULE_050_PROC_VIEW WITH PARAMETERS ('placeholder' = ('$$i_tenant_id$$', '''0001'''));

Next step is, I try to call this from within a (different) SQL-procedure. I declare the following:

...

declare cursor c_cursor for select * from NEO_CCG502K10CQSC8NX8DNT6J7MA.RULE_050_PROC_VIEW WITH PARAMETERS ('placeholder' = ('$$i_tenant_id$$', '''0001'''));

...

The procedure is compilable, and I can create it, but once I execute it, I get the following error:

Could not execute 'call NEO_CCG502K10CQSC8NX8DNT6J7MA.EVALUATE_RULES_050_PROC_V1( '0001', '0001', o_result =>? )' in 475 ms 994 µs .

SAP DBTech JDBC: [257]: sql syntax error:  [257] "NEO_CCG502K10CQSC8NX8DNT6J7MA"."EVALUATE_RULES_050_PROC_V1": line 23 col 2 (at pos 873): [257] (range 3) sql syntax error exception: sql syntax error: incorrect syntax near "0001": line 1 col 126 (at pos 126)

I then tried to play around with the Input Variable, for instance this

select * from NEO_CCG502K10CQSC8NX8DNT6J7MA.RULE_050_PROC_VIEW WITH PARAMETERS ('placeholder' = ('$$i_tenant_id$$', '0001'));

works as well, once I call it directly from the SQL console, but once I use exactly this SQL syntax in my procedure,

then it gives me the following error:

Could not execute 'call NEO_CCG502K10CQSC8NX8DNT6J7MA.EVALUATE_RULES_050_PROC_V1( '0001', '0001', o_result =>? )' in 304 ms 184 µs .

SAP DBTech JDBC: [1301]: numeric or value error:  [1301] "NEO_CCG502K10CQSC8NX8DNT6J7MA"."EVALUATE_RULES_050_PROC_V1": line 25 col 3 (at pos 946): [1301] (range 3) numeric or value error exception

Once to try to call the result view from my SQL console like this:

select * from NEO_CCG502K10CQSC8NX8DNT6J7MA.RULE_050_PROC_VIEW WITH PARAMETERS ('placeholder' = ('$$i_tenant_id$$', '"0001"'));

then I get the following error:

Could not execute 'select * from NEO_CCG502K10CQSC8NX8DNT6J7MA.RULE_050_PROC_VIEW WITH PARAMETERS ('placeholder' = ...' in 282 ms 621 µs .

SAP DBTech JDBC: [2048]: column store error: search table error:  [34023] Internal error during instantiating calculation model

What am I missing? Can anyone give me some advice?