HTTP Handler strange behavior when updating HR data

Hi there,
Brief summary: Executing a function module from inside a HTTP Handler gives a different result than executing the function module direcly through SE37. The user defined for the SICF service calling the Handler is the same as the one we use for SE37.
Details:
We have created an HTTP Handler implementing the interface IF_HTTP_EXTENSION.
We have also defined a new service through SICF that access the previous Handler.
We have a web application accessing SAP through that HTTP Handler.
Within our HTTP Handler (within the IF_HTTP_EXTENSION~HANDLE_REQUEST method), we are calling a custom function module that perform different actions depending on the parameters received.
The solution works really well but we have identified an issue for some actions: when we execute some background updates (like calling HR_INFOTYPE_OPERATION) and the transaction returns an error, the HTTP Handler suddenly stops and the HTTP call returns an error.
The strange thing is that executing the function module directly (without calling the HTTP handler from the web) captures the error and continues the execution of the function module.
It seems that when the function module gets called from the Handler, there is some screen or user parameters that affect the result of the transaction, and we cannot find the way to capture that error and let the Handler to continue.
Any idea, suggestion, alternative solution would be much appreciated
Best regards,
Miguel

Hi Olivier, thanks for your reply, nice to see you are still out there
You raise a good point. We are reusing an "old" web app that was connecting to SAP through the DCOM connector. The RFC (the function module I mentioned on my first post) cannot be modified as it handles tons of different web calls, so we decided to call that RFC from the Handler.
It may not be the best solution, but the web app has been modified to call the different handlers (for the different environments) and still hope to find a work around our current problem.
Other than that, it works really well and performance is amazing
Thanks again,
Miguel

Similar Messages

  • Strange behavior when I try to match a mpeg2 video with the menu size image

    hello all,
    got a strange behavior when I try to match a mpeg2 video with the menu size image,
    menu size is 1920 x 1080 (photoshop file)
    mpeg2 video 1920 x 1080 (1second 19 frames)
    the idea is to go from the menu link to the next sub menu with a video as transition, using the same image on the video as the menu and then zooming in (in my case through a door) to end up after the zoom in the next level menu.
    I try to achieve a smooth change from the first play menu picture to the video image into the next level menu picture,
    everything works good but when I watch it on the preview mode on the left and right side of the video it shows a black bar at least 15 to 25 pixel wide, that means the video image appears smaller and it causes a little jump what is preety disturbing and can not be sold as a professional work.
    why encore treats the video size and picture differently?
    it's gratful appreciated if somebody who got experience in a transition I mentioned obove posting any solution,
    thankl you all

    Hi all
    I'm having a very similar problem with Encore. My menu items are distorting (a minor but infuriating squashing top and bottom) when previewed.
    Using Encore,PS,AE CS4
    DVD format SD PAL DV widescreen
    I created 3 menu frameworks and 3 video transistions to link them in AE. On the final frame of each transistion in AE I saved the frame with Composition>Save Frame to> Photoshop Layers. In PS i added the button functionality and saved.
    In Encore I use Dynamic link to import AE transistions as timelines and imported each PS menu twice, once as "menu" and once as an asset.
    When I link the timeline transistions to the menu and preview at the point the menu begins the whole image is squashed. When the next transistion is activated the image returns to its original size.
    Thinking it was the use of PS that was causing the menus to distort I used the menu asset directly in the time line. Obviously no menu functionality but also no distortion. The AE transistion flowed straight into the "menu" just as I expected.
    In the properties panel the menu Aspect Ratio 16:9
    The AE transistions PAR is SD PALwidescreen 1.4587.
    I did try the Blu Ray suggestion above but the same distortion was apparent.
    Any suggestions short of squashing the video to match the menu would be welcomed.
    Thanks

  • Strange behavior when performing query

    Hi all,
    I'm getting some odd behavior when doing different kinds of queries and am hoping someone here can help out. I have a database with continuously updated PDML data (http://www.nbee.org/doku.php?id=netpdl:pdml_specification). I am opening anf querying the database from a seperate application. This appears to work fine for certain types of queries, and fails for other types with error messages that don't seem relevant. Specifically, some queries that use the attribute axis as the last leaf seem to get fouled up. Here are some example queries and the results I get:
    First some queries that work:
    Query: "string(collection()[1]/packet[1]/proto[1]/field[1]/@size)"
    Result:
    A valid number
    Query: "collection()//proto[@name='geninfo']/field[@name='len']"
    Result:
    <field name="len" pos="0" show="64" showname="Packet Length" value="40" size="64"/>
    <field name="len" pos="0" show="64" showname="Packet Length" value="40" size="64"/>
    <field name="len" pos="0" show="64" showname="Packet Length" value="40" size="64"/>
    Now some queries that don't work:
    Query: "collection()//packet/proto/field/@size"
    Result:
    Output to stdout - "Transaction specified for a non-transactional database"
    and an Exception with message - "Error: Invalid argument"
    Query: "collection()//proto[@name='geninfo']/field[@name='len']/@name" (notice it's almost identical to above working query except with @name)
    Result:
    Same as above, an stdout about transactions and an exception about invalid arguments
    What I find most interesting is that both errors (the one printed to stdout and the one I get in my exception) don't seem to have anything to do with the problem. I know the transactional stuff isn't messed up because the exact same code works for certain queries. I'm not sure about the "Error: Invalid argument" exception, but it's not very helpful by itself. I'm using the Java library and DBXML version 2.4.16. Anyone have any ideas what's going on?
    Thanks,
    Dave

    As a follow-up, it appears that I just needed to wrap my attributes in the XPath string() function. For example "collection()//packet/proto/field/@size" does not work while "collection()//packet/proto/field/string(@size)" works fine.
    I'm still not sure why it's spitting out an error message to stdout about transactions not being enabled when the only real error is with the query, but at least it's working now.

  • Strange behavior in updating and standbying

    Hi,
    In the last 4 weeks I've been facing some strange behavior from my iPhone 4. First, sometimes when I pick it up its screen is already on (dimmed) and showing the lock screen (with some elements missing though). Secondly, and more annoying yet, when I try to update the apps via app store the initial process occurs normally but after loading the first app it begins installing it and keeps in it forever. I have to soft reset the phone. After that, the app (that kept installing) seems ok. So, I repeat the process of update from the beginning (app after app).
    I've already tried to hard reset the device without any luck with any of the problems. Any suggestions about what is going on?
    Thanks
    Adriano

    Hi,
    I'm not clear on the distinction you're making for type of access.
    Well, to be honest, now that I've re-read that this morning I'm not completely sure what I had in mind either. I think I was thinking about using the collection in what I believe the documentation calls "slices". However, given the other information you've posted it doesn't seem like that would be a feasible alternative in any case.
    I understand what you mean about varray and nested-table with the upper-bound limit for the varray's... I've not actually compared the two for performance so my idea that there may be a difference could be completely incorrect. Beyond that, creating code that only works in batches (or slices) may be less performant than what you currently have.
    I hope I've not detracted from the thread and maybe someone else with more experience will have an "ah ha!" sort of observation.
    Regards,
    Mark

  • Strange behavior in Webi : No data returned

    Hi,
    I am experiencing a strange behavior.
    I have created a condition object in universe.
    Start Date : DATE_DIM_TBL.DAY >= UPPER(to_char(to_date(@prompt('Enter start date','D',,mono,free),'dd/mm/yyyy HH24:MI:SS'),'dd-mon-yy'))
    End Date : DATE_DIM_TBL.DAY <= UPPER(to_char(to_date(@prompt('Enter end date','D',,mono,free),'dd/mm/yyyy HH24:MI:SS'),'dd-mon-yy'))
    Now when I create report to display years between 01-05-2005 and 04-05-2009. Report does not return any data.
    Where as when I execute the generated query from SQL prompt it return correct data.
    SELECT
      DATE_DIM_TBL.CALENDAR_YEAR_NAME
    FROM
      DATE_DIM_TBL
    WHERE
       ( DATE_DIM_TBL.DAY >= UPPER(to_char(to_date('01-05-2005 00:00:00','dd/mm/yyyy HH24:MI:SS'),'dd-mon-yy'))  )
       AND
       ( DATE_DIM_TBL.DAY <= UPPER(to_char(to_date('04-05-2009 00:00:00','dd/mm/yyyy HH24:MI:SS'),'dd-mon-yy'))  )
    can somebody please give any pointers about what has gone wrong in WebI.
    --Kuldeep

    Please find below
    20-APR-05
    02-JAN-00
    03-JAN-00
    04-JAN-07
    05-JAN-00
    06-JAN-08
    07-JAN-00
    08-JAN-09
    09-JAN-11
    Yes I tried SQL*PLus on BOBJ Server machine. SQL returns correct data.
    Do you think is it some setting in WebI, As It hapopens only when I use condition object as a filter in Webi report. When I create Report filter it works fine.
    --Kuldeep

  • Error when updating the data from DSO to cube

    Hi,
    I am getting the error when uploading the data from the ods to cube.
    The following is the error message.
    Unable to determine period for date 20090101, fiscal year variant Z2: Error #
    How can i solve this issue.
    Regards
    Annie

    Hi ,
    fiscal year variant, go into Customizing for Financial Accounting (FI) under Financial Accounting Global Settings >>>Fiscal Year >>>Maintain Fiscal Year Variant.
    check this link ..
    http://help.sap.com/saphelp_scm41/helpdata/en/50/0d89f2ad919c40b95b9ae7583c8c96/frameset.htm
    http://help.sap.com/saphelp_scm41/helpdata/en/50/0d89f2ad919c40b95b9ae7583c8c96/content.htm
    Regards,
    shikha

  • Password change issue when updating user data in SAP ABAP system

    Hi Guru's,
    One of my reconciliation tasks part of the reconciliation job I've created is doing some strange password updates.
    As you can see below the task selects all users part of my identity store that are part of the account attribute of the particular ABAP system.
    Once these users are selected the task updates different data like username, validto, ... but the task is updating a lot of other things that are not part of the destination tab. What is causing the biggest issue is the password fields that are updated in the ABAP system like, password, productive password, ...
    Can you please advise if I missed something and how to solve?
    Thanks a lot,
    Laurent

    Hello Steffi,
    Yes in the ABAP systems they have the same timestamp. No other jobs are running at the same time.
    It is only happening to a few users depending on the ABAP system. On some ABAP systems there are only a few users for which the PW is reset and other systems 300.
    Example below of a system where I updated all user. In my pass only the following attributes should have been pushed thru to the ABAP system.
    However the valid from, accounting number and password have been updated as well.
    Thx,
    Laurent

  • Strange behavior when compiled with Visual C++ 2005 Express Edition

    I wrote a program for Berkeley DBXML 2.2.13 on Windows XP that executes a query 5 times and prints out the runtime for each execution. When I run the program using the Debug version of the libraries the queries take about the same time to execute for all 5 executions. When I run the program using the Release version of the libraries the execution time of the queries degrades significantly between each execution. One query takes 2 seconds to execute the first time and 20 to execute the last. I have tried recompiling the Release version without optimization and the strange behavior persists. I have run both the Release and Debug versions several times and gotten the same results, so I doubt it is a program running in the background that is slowing the execution.
    This behavior does not happen when I compile the program under Unbuntu using g++, so I assume it is caused by some flag set in the Release version of Visual C++. Has anyone else experienced this behavior and know what flags I should or should not set?
    Thanks for your time.
    Lauren

    Hi Lauren,
    We've had users experience this before, and it seems to be a bug/"feature" of the default windows heap allocator. Please see this post for a bit more explanation and a solution:
    Re: Performance Issue caused by XmlResults.next().asString();
    John

  • Strange behavior when shutting down - only works once

    Hi there,
    I'm taking advice from around the forum to create a shutdown button for my VI. When the user sends the shutdown command, the loop stops all processes running and turns off all heaters. Then when this is complete it passes a shutdown variable to all loops allowing them to close. This seems to work perfectly once, but when I hit run again and test it again, it fails. If I abort the VI and restart again, the shutdown works. I can't quite figure out what the behavior is doing here.
    Attached is my program, it's a little messy but everything works, and it has to be added to by someone who isn't me, so I tried to keep everything clear and seperate. If anyone's got any ideas as to what is causing this strange behavior I'd love to hear it.
    Thanks,
    Chris.
    Attachments:
    Main VI.vi ‏259 KB

    First of all, your event loop should look something more like this.  Otherwise you won't exit the loop until another event comes through.
    Secondly, I think you are writing to the wrong control here (this is in your timed loop).  I have no clue what is in the subVI, but I don't think it should be looking at the Shutdown PB control.  That control should be in it's event case.  But here you are trying to cause the shutdown event, but you are writing to the wrong control.
    There are only two ways to tell somebody thanks: Kudos and Marked Solutions
    Unofficial Forum Rules and Guidelines
    Attachments:
    Event Structure issue.png ‏10 KB
    wrong property node.png ‏3 KB

  • Strange behavior when "trying" to burn a disk.

    At least I think this is strange behavior, following the instructions in the help viewer, topic "Backing up your music to a CD or DVD." One thing for sure, it's not burning a disc and it ain't telling me why.
    I select the playlist and click "Burn Disc"
    Requests a blank disc. I insert one.
    Message "checking media"
    after about a minute, the disc ejects
    Message "checking media" displays another 30 seconds
    No other messages displayed.
    What the ??? Shouldn't it at least display an error message??
    Super-Drive information..... from System Profiler
    MATSHITA DVD-R UJ-845C:
    Firmware Revision: DPP9
    Interconnect: ATAPI
    Burn Support: Yes (Apple Shipped/Supported)
    Cache: 2048 KB
    Reads DVD: Yes
    CD-Write: -R, -RW
    DVD-Write: -R, -RW, +R, +RW
    Burn Underrun Protection CD: Yes
    Burn Underrun Protection DVD: Yes
    Write Strategies: CD-TAO, CD-SAO, DVD-DAO
    Media: No
    Using Sony DVD+R discs.
    Mac Mini   Mac OS X (10.4.8)  

    Whoops, sorry, this was while burning a playlist.
    I've also tried some Pleomax CD disks (Samsung), with no luck. I've tried burning directly from iTunes, and from Toast 7.
    Of the types I've tried, the Sony DVD's have been the best for me. Haven't tried many types though. Maybe I can get disks one at a time until I find a brand my burner likes.
    Still, isn't it strange that there is NO error message? It just quits?
    Mac Mini   Mac OS X (10.4.8)  

  • Strange behavior when resizing with 'Selection Tool'

    Hello,
    I have a really strange issue when using selection tool to resize object. I have several rectangles stacked and grouped inside a layer. When I select one of them (either in Layer panel or with Group selection tool) and switch to 'Selection Tool (V)' to resize it, this tool behaves diferently at diferent zoom levels. Only at 200% zoom it behaves as it shoul.. With 100-150% zoom result is diferent - like there's a invisible box around selected object and I'm resizing it.. I've tried to explain the issue in attached image.. Anyone knows why is this happening?
    Thanks.

    - make new document, size 16383x16383px
    - create rectangle any width, hight say 200px (or anything less then 320px for 5% zoom)
    - select rectangle and show bounding box
    - try resizing hight by stretching bounding box
    - this is how same selection looks at diferent zoom levels
    - same zoom level as befor, hight 100px..

  • Strange behavior when using Labview to collect data from Tektronix tds8200 oscillosco​pe

    I have hit a wall in trying to figure this one out. The problem I am having is that my program does not start the oscilloscope when it should.
    I am using a Tektronix TDS8200 oscilloscope. My goal is to collect waveform data from the oscilloscope using Labview. My program first initializes and configures the oscilloscope; this part of the program runs fine.
    The second part of the program begins the data aquisition using the  'tktds8k Start or Stop Aquisitions.vi' function, which is equivalent to pressing the Run button on the scope. The 'tktds8k Get Waveform.vi' function is then used and should ideally return the data, which I have connected to a waveform graph for visualization.
    When I run my program, the first part executes without issue, but as soon as the program gets to the Get Waveform function, the Run button on the scope, which is green when running, turns off; the program then times out, and no data is collected.
    Here's where it gets weird. I went through some debugging to try and figure this out, and I put breakpoints on both the Start and Get Waveform functions so that I could step through the later part of the program. The program continues through the Start function, and the Run button on the scope is green. The breakpoint for the Get Waveform function is reached, and when I press continue, the Run button turns off and then turns back on almost immediately; the data is collected, the waveform graph is displayed, and the program exits without an error.
    I thought timing might be the issue, so I made the program wait as long as 5 seconds between the Start and Get Waveform functions, and that did not work. I also tried moving the Start function to before the configuration functions, and removing the Start function altogether; neither method worked.
    Are there any thoughts on why the program works when I have the breakpoints enabled and doesn't when the breakpoints are disabled? I am sure that there is an easy fix, but I haven't been able to find a solution.
    I have attached a pdf containg information on the Oscilloscope functions (tktds8k.pdf), and I have also attached my program.
    Solved!
    Go to Solution.
    Attachments:
    tktds8k.pdf ‏1424 KB
    set_up_osc.vi ‏32 KB

    Thank you for the swift replies.
    After Bill asked about the ID Query, I decided to try a few things that I had already gone over just to double check. I was suspspicious that timing was the source of the error. Using a timed while loop, I set the wait time to 10 seconds and got results. As it turns out, 5 seconds is not long enough, but 8 seconds is a sufficient wait time for the program to work. With the problem solved, I am still mystified as to why waiting 8 seconds is required.
    I think that the oscilloscope must be given sufficient time to display the signal on-screen before the Start Aquistion or Get Waveform functions are used. With this logic in mind, the breakpoints were acting as a sort of wait, allowing the signal to be displayed before continuing through the program.
    In response to Jeff, I am indeed using an external direct trigger. The hardware is sound, but apparently, my coding could use some work.
    I have attached the modified code. I am certain that there is a more elegant solution to the timing than simply slapping a timed while loop on the code. Any suggestions?

  • Strange behavior when getting data and storing it into a tdms file

    Hi
    I am getting data from a VISA device(10bytes every 20ms) in a string buffer and then using indexing array get each element which is 2 bytes , filtering the data in real time using a butterworth filter and then storing the data to a TDMS.
    The problem is that the data goes haywire after a brief the the different elemetns just switch. So for example x becomes resistance, GSR becomes  etc
    I have uploaded the VI and the TDMS file converted to excel format

    You really should be writing to the TDMS file while you acquire the data.  By using the Autoindexing Tunnels to build the arrays, you are causing A LOT of memory allocations, which causes things to really slow down.  I would venture to say that you are missing data, causing what looks like a shift in the data.  But you would need to be looking at the error coming out of the VISA Read to know for sure.
    So look into the Producer/Consumer.  The idea is to use a parallel loop to log the data while your current loop reads and processes the data.  The data is sent to the consumer loop via a queue.

  • Strange behavior when pressing ENTER in a form page

    Hi there
    Please go to the address:
    http://apex.oracle.com/pls/otn/f?p=20104:4
    If you click the Create button, enter anything in the Name field and press ENTER while in the field, you will be taken back to the main page, the message Action Processed will be displayed, but no data will be inserted. If you click on the Create button though, it works fine.
    Have anyone seen this? Any explanation for it? How to fix it? I could not find any reason for this behavior (the Name field is NOT an always submit page when enter pressed field type).
    (btw the two pages are very simple, based on wizards without any customisation)
    Thanks
    Luis

    i guess, trying the input form, the input text field for Name is an item 'textfield submit on enter'. But what happens is that when enter is pressed no additionally submit information is send, which is the case with buttons. Buttons call the doSubmit('<BUTTONNAME>') javascript function on click, allowing the process to run when thee are set conditionally.
    Carl's answer by setting the process unconditional will also result this action will take place when another buton (without direct redirect) will fire this.).
    So additionally when multple buttons and processes some new javascript kan be used to do the right submit on Enter. Example:
    In Page header, or shared javascript import file create a JS function submitOnEnter
    function submitOnEnter(submit,e){
    var keycode;
    if (window.event) keycode = window.event.keyCode;
    else if (e) keycode = e.which;
    else return true;
    if (keycode == 13){
    doSubmit(submit);
    return false;
    }else{
    return true;
    Then with you're item you define the js-event onKeyPress to call the function submitOnEnter
    onkeypress="javascript:return submitOnEnter('CREATE',event);"

  • Strange behavior when trying to change backgrounds after Aperture install

    Running 10.5.1 on my MacBook Pro without a hitch until today. Installed Aperture 1.5 and updated it to 1.5.6 last night and it too works fine.
    However, today when I went to change my background image in System Preferences I encountered some strange voodoo. I originally under the Desktop applet in System Preferences I had the default images and folders supplied with the operating system, and one images folder I maintain under my Documents folder for my personal backgrounds.
    Now I am no longer able to view my personal backgrounds folder in the System Preferences Desktop applet to select one of my images as a background. Instead I have, literally, two occurrences of the Aperture Projects folder listed under the default operating system background image options.
    Initially I found I cannot discard either instance of the Aperture folder, nor can I re-add my personal images folder (states the folder is already in the list as an option). Naturally I figured something might be amiss with permissions. So I rebooted and ran the Disk Utility which checked my drive to find no pertinent errors. No change in symptom.
    Next I renamed my personal images folder. This produced a change in symptom.
    Now I only observed one occurrence of the Aperture Projects Folder in the System Preferences applet. However, when I went to add my newly renamed personal images folder the second occurrence of the Aperture Projects folder appeared and would not let me access anything therein.
    Could certainly be a problem specific to 10.5.1, but thought it worth posting here in case anybody else wants to try to troubleshoot or has the same voodoo. Plus, the problem did not occur in 10.4.X running Aperture - so I think it is pretty specific to 10.5.1.
    Not a show stopper, but certainly weird. Hopefully not indicative of any additional problems lurking under the covers.
    TIA for any input or responses.

    it is a bug in leopard that we have uncovered ...
    adding an iphoto library solves the issue ...
    http://discussions.apple.com/thread.jspa?messageID=5847357&#5847357

Maybe you are looking for

  • Can't update to IOS 5.1, says IOS 5.0.1 is the latest software, Can't update to IOS 5.1, says IOS 5.0.1 is the latest software

    Since a few hours ago, IOS 5.1 has been released but on both my iPad and my iPhone, it still says that my software is up to date though I have IOS 5.0.1. Both my parents have been able to update, and most of my friends are updating right now. Any hel

  • Pictures not being saved to camera roll (or anywhere)

    Hi all, I was recently having some issues with sending/receiving data over 3G on my iP4. To try & resolve this, I did a clean restore of my phone. Since then, I have been unable to save any images from the camera (or email, mms etc) to my phone. Init

  • Changes in customer master data

    Hi Is there a different way (apart from using IDOCS) to identify changes (create/update/delete)  in customer's master data to the following  Condition Class: BPAR TESTRF DEBI KLIM Best Regards Tal

  • J2EE SDK client-server on my PC

    dear friends, I need your help, please. First I studied J2SE as client on my PC. Now I should like to study J2EE client-server, but I own only one PC. Is it possible to simulate a J2EE client-server situation inside a single PC? And how? many many th

  • Print a JFrame in a PDF file ?

    hello, I just downloaded the iText library which allows the generation of pdf files. i'm just wondering if there is a way to print a JFrame into a pdf File, in other words I want that the Jframe which holds all my components to be printed in a pdf fi