Release Notes: BIQ 5.02

Install (general)

The BIQ install has changed to an msi-based InstallShield install. There are three types of BIQ install kit. The first part of the filename indicates the type:

BIQSetup / Full install. Installs the BIQ client, installs and starts BIQ services. Can be used on machines that have never had BIQ installed or to upgrade existing BIQ version 3.2+ full install. Note: full installs are refused if a client install is present.
BIQUpgradeSetup / Upgrade install. Upgrades the BIQ client and server. Can only be used on machines that have BIQ version 3.2+ full installed.
BIQClientSetup / Client-only. Installs the BIQ client. Note: client installs are refused if a full install is present.

The second part of the filename is the version. For example, an upgrade-only install kit for version 4.01 revision 5000 would be named: BIQUpgradeSetup_4_01_5000.exe

When a new version is released, Upgrade and Client-only kits can obtained at Full installs require a userid/password login to a different location, supplied by BIQ or by a BIQ distributor.

Administrator privileges are required to run any of these kits. Although some forms of install (client-only, upgrade-only) were possible in many cases under Windows XP without administrator privilege, changes made by Microsoft to the Windows security model under Vista and Windows 7, and/or by company IT departments under XP, now make this impossible in most cases.

New Installs

Run the install kit and follow these steps:

  1. Accept the terms in the license agreement.
  2. Enter the Serial Number.
  3. Choose the destination folder. In most cases the default folder is the correct one to use.
  4. (Full only) Choose the data folder where datasets and log files will be stored. The default is the windows [CommonAppDataFolder] which is usually the correct choice.
  5. (Full only) Select the Port Number that the BIQEngine and MySQL services will use. The defaults is 80 and 3306 respectively. If either of these ports are busyyou can enter a new one, or have the installer find free ports by hitting the Auto Select button. Note: If you plan to install Skype or some other program which also uses port 80 you should change the BIQEngine port now.
  6. Press Install.

Note: If MySQL is already installed on the machine, and it is not a BIQ install of MySQL, the install process will terminate with an error.

Upgrade Installs

When upgrading from BIQ version 3.283 or earlier, the install works similarly to a new install, except that existing datasets are preserved (there is nothing to preserve in client-only upgrades). Please see the New Installs section above.

Note: Release 3.283 used a data directory that is only writeable by the user that installed the application. If this BIQ install will be used by multiple users on the same machine, please change the data folder to a common area during the install.

When upgrading BIQ version 4.0 or greater to a new version of BIQ, the installer does not need to ask any questions. It will stop any services, upgrade the application, and restart the services.

Install Issues

If there are any issues during the install, please follow these steps to help diagnose the problem: (1) Check the "Show the Windows Installer log" checkbox before pressing Finish. This will bring up BIQ_Install.log in a text editor. Save this file to a place you can remember, and then please email the file to .

If you have already closed the installer, the log file can be found in: %TEMP%\BIQ_Install.log. To access this folder, type the string “%TEMP%” into Windows Explorer (NOT Internet Explorer, Windows Explorer – the application that launches when you click the “My Computer” icon). Omit the double quotes.

Release 5.02.xxxx (changes since 5.0.1.xxxx)

Significant changes

  1. Performance improvements: Substantial performance improvements were made in numerous areas of the product, including:
  • Searching by RuleID (issue #3707)
  • Starting the viewer with datasets containing large numbers of nodes (issue #3693)
  • Dimension editor loading and scrolling (issue #3665)
  • Selecting a dataset for modification in the loader (issues #3687, #3632)
  • Change file importing (issue #3691)
  • Transaction data extract (issue #3603)
  • Searching rules, expand all (issue #3615)
  • Create sub dataset (issue #3622)
  • Drilling, repainting data blocks (Issue #3627)
  • Rule application (red and green buttons) (issue #3698)
  1. New feature: It is now possible to import multiple change files in one step (issue #3633).
  2. New feature: It is now possible to import and export filter sets (issue #3634).
  3. New feature: It is now possible to create time dimension containing dates beyond 2038 (issue #3637).
  4. New feature: It is now possible to invert rule marks (issue #3640).
  5. New feature: It is now possible to create a dimension derived from mapping rules (issue #3657).
  6. New feature: Two new rules reports have been added: Inconsistent and duplicate rules (issue #3659).
  7. New feature: Greatly expanded rule filtering in rule management window (issue #3663).
  8. New feature: Disable one or more rule groups to gauge impact on dataset (issue #3692).
  9. New feature:Allow editing of target and filter node within any number of marked rules (issue #3674).
  10. New feature: Significant changes to the dataset are now written to anaudit log(issue #2651).
  11. New feature: A new template was added for "Years, months" to time dimension node generator (issue #3683).
  12. New feature: New installations on Windows 10 machines will now default to square window corners. A new preference has been added that overrides the default (issue #3651).
  13. New feature: The fields tab in the loader now contains more descriptive field count information (issue #3631).
  14. New feature:Allow the user to get to recent datasets and other tools by right clicking the viewer icon on the start bar. Also, progress of the current task is now displayed within the icon located on the start bar (issue #3656).

Viewer, Loader, Chopper

  1. The “buffer and memory management” functions on the server admin tab have been moved to a separate tab (issue #3678).
  2. The exception handling dialog has been improved and now allows “Copy to clipboard” and “Mail to BIQ” (issue #2357).
  3. Drop down combo boxed throughout the entire system now allow for searching using the first letter by pressing upper case letters on the keyboard (issue #3653).
  4. The splash screens in the dataset loader and chopper were not resizing correctly (issue #3662).
  5. Spurious delays have been removed from the license information window (issue #3625).
  6. Moving the mouse away from the primary monitor while BIQ was initializing could cause resizing and window location issues (issue #3626).

Viewer

  1. It was possible to resize a data block to a size that was too small to display correctly and this caused errors (issue #3695).
  2. A .net compatibility issue with Windows Server 2003 was addressed (issue #3706).
  3. Incorrectly formatted rules (rare) now display an error format that includes the missing NodeID (issue #3708).
  4. Data blocks in the viewer now have a minimum size of at least 2 data lines (issue #3666).
  5. The dropdown on the nodal CML editing window was not working correctly (issue #3667).
  6. Application fault when renaming measures from the fields tab, related to cross checking meta measures (issue #3675).
  7. The logic that controls the “don’t ask me again” sequence was broken in the import rules function (issues #3239, #3672).
  8. Create sub dataset caused an application fault in same cases (issue #3619).
  9. Incorrect preference handling when only non-dimension blocks are displayed (issue #3620).
  10. Incorrect block size after cloning dimension (issue #3621).
  11. Minor adjustment to the show filters combo box size (issue #3624).
  12. In some cases, multi-dimensional extracts with reference queries at drill positions not showing accurate results (issue #3641).
  13. Progress bar not working correctly for extract containing less than 1000 transactions (issue #3543).
  14. Cloning a dimension with rules changes the order of the rules in the cloned dimension, also LastModDate is reset to all zeros (issue #3654).
  15. BuildCMLFastCompare entry dialog auto selecting and not displayed for user to choose a filter (issue #3655).
  16. In rare circumstances, charting could cause an application fault on datasets in need of compaction (issue #3660).
  17. The “Configure meta extract” window was resized to handle display resolutions less than 1300 x 958 (issue #3661).

Loader

  1. The loader allowed duplicate field names when adding fields to fact (issue #3715).
  2. The dimension node report sometimes caused a fault (Issue #3717).
  3. Dataset sizing information was not being updated after removing a transaction source (issue #3699).
  4. Thin index could potentially cause a fault (issue #3696).
  5. In rare circumstances, the loader’s main window could be displayed off of the primary display (issue #3681).
  6. Editing an expression for field types other than string could cause a fault (issue #3682).
  7. Dates fields saved during transaction browsing contained the period character (issue #3686).
  8. The truncate before load option was not always correctly recorded for fact and index sources (issue #3688).
  9. It is now possible to add a node to an existing time or range node when creating these types of dimensions (issue #3690).
  10. The loader main window now contains notifications (issue #3676).
  11. It is no longer possible to name a field “TXID” or “Count” (issue #3677).
  12. The loader could skip the first line of an index or transaction source if the input file did not contain a header line (issue #3679).
  13. Additional dataset corruption checks have been added to the loader before the dataset can be modified (issue #3703).
  14. In rare cases, a fault could occur when deriving a flattened dimension when an expression field as missing from one or more source (issue #3608).
  15. The loader could potentially drop the demo dataset because of suspected damage (issue #3611).
  16. The loader would sometimes display an extra icon in the taskbar (issue #3635).
  17. When adding multiple fields to the source definition from a file the names are all the same as the first field name (issue #3630, L0040).
  18. Add the ability when browsing transactions to save without dynamic and/or TXID (Issue #3711).
  19. When deleting one dimension in the new loader if there was an extract conflict, the dimension is no longer deleted (issue #3648).

Chopper

  1. The chopper no longer shows the TXID field when it is run integrated into the loader (issue #3673).
  2. The chopper can now load transforms created by older versions of the chopper (issue #3680).
  3. The chopper no longer displays erroneous “Possible syntax error” warnings (issue #3697).
  4. Changes to the separation character are now handled correctly by the chopper (issue #3702).

Admin

  1. The SQL tool located in the administration utility tool now works correctly (issue #3639).

Release 5.01.xxxx (changes since 5.0.9440)

Significant changes

  1. New feature: The loader user interface has been completely overhauled with a focus on usability and clarity (issue #3350). In some cases it may be necessary to use the legacy dataset loader tool to work around a bug or missing feature in the new loader. To launch the legacy loader from the viewer hold down the shift key when clicking on the Dataset menu in the upper left. This will show a menu option called 'Legacy dataset loader' which gives the option to open the legacy loader in either the main screen or to modify the current dataset.
  2. New feature: The File Transform Utility (chopper) user interface has been overhauled with a focus on usability and clarity (issue #3517).
  3. BIQ now requires version 4.0 of the .NET Framework (issues #3356 and #3531). For most systems this will be handled automatically during the upgrade process.

Viewer, Loader, Chopper

  1. It is now possible to take a snapshot of an existing dataset before making major changes. The snapshot can be restored to revert to the previous dataset state (issue #3270).
  2. The License Information dialog was improved and a 'mail to BIQ' button was added (issue #3527).
  3. Upgrade SftTree libraries for Windows 8.1 (issue #3530).
  4. Slay viewer appears not to work correctly (issue #3597).
  5. Splash screen displaying a blank form (issue #3599).
  6. In some cases updating the license installation code quietly failed for non-administrator users (issue #3605).

Viewer

  1. New feature: Allow renaming of view names (issue #3519).
  2. Extract with hidden dimension, automatic extract item, and usertries to make a pivot table and cannot (issue #3480).
  3. Saving a report with a name with quote in it causes a fault (issue #3482).
  4. Charts within Excel are now forced to update after a table is filled (issue #3485).
  5. Having dimensions hidden when using pivot walkthrough could cause fault (issue #3492).
  6. In rare cases, connections to the engine from clients would be blocked (issue #3493).
  7. Dataset selection window does not follow viewer main window in some cases (issue #3506).
  8. Add the ability to ignore non-fatal errors when importing extracts (issue #3514).
  9. At the time of reloading windows, map measure window is partially visible for few seconds (issue #3518).
  10. Dimension block menus did not remain visible ifthe pink mapping window was open (issue #3526).
  11. Renaming a measure can sometimes cause application hang when CML measures are also present (issue #3528).
  12. Importing rules files with blank comments can cause fault. Older rules exports without comments can cause fault (issue #3529).
  13. When creating dimensional measures on dimensions with a large number of nodes, a fault could occur (issue #3532).
  14. Transactional extract sources and transaction extract definitions were not being updated when renaming dimensions and indexes (issue #3537).
  15. Several buttons on the f(x) window and one of the helper drop downs on the TCML window were nonfunctional (issue #3538).
  16. 64 bit transactional extracts to csv file may sometimes produce a fault (issue #3541).
  17. Zeros can occur in CML output if the CML code contains a variable with an asterisk (issue #3542).
  18. Tabs were not being allowed in the CML code text box (issue #3543).
  19. Complex transactional extracts with many indexed fields can sometime return N/A rather than the actual field contents (issue #3544).
  20. Drag drop on transaction extract definition window did not work correctly and could sometimes fault (issue #3545).
  21. Viewing items by percentage of item no longer working (issue #3546).
  22. Expand/collapse all shows wrong status message in progress bar (issue #3550).
  23. The Stddev statisic can sometimes cause the engine to fault (issue #3553).
  24. Make mysql index addition optional when importing rules (mapping window). Checked by default (issue #3554).
  25. In non-artificial baseline changes files, moving a child level node to top level did not work correctly (issue #3562).
  26. "Show all nodes" option was did not work in all cases (issue #3563).
  27. Lock the dimension editor button to prevent double entry (issue #3568).
  28. Viewer labels were not remembering font size (issue #3570).
  29. Don't ask the user "Are you sure" when doing Loader-Modify Current (issue #3571).
  30. If the license validation takes a long time then the timer for running it gets kicked off twice which causes an exception (issue #3572).
  31. Labels should remain on top at all times (issue #3573).
  32. Phrase mapper window: measure selection combo box not working (issue #3574).
  33. When searching for very long strings it was possible to get "Error from DLL_MatchResultsGetAll: -3 (issue #3575)"
  34. Rules remap (remove dimension) performance was improved (issue #3576).
  35. Asking engine to put two files within any second would cause the file to be overwritten (issue #3580).
  36. Joining popped out blocks on exit would crash if panel was already destroyed (issue #3581).
  37. Fix a problem with MySQL disconnects and deleted databases being selected from the last MySQL pool connection (issue #3582).
  38. Drag and drop does not work on Windows XP x64 Professional on the extract wizard form (issue #3585).
  39. Deleting hidden dimension from within the viewer causes a fault (issue #3587).
  40. Importing/exporting rules from the rules editor screen could sometimes cause a fault (issue #3589).
  41. Unknown dimension name in dimension preferences will cause the viewer to spin forever (issue #3594).
  42. Warn user that tracking may not be enabled if there are no last effecting rules (issue #3595).
  43. Maximize causes the taskbar to be obscured (issue #3596).
  44. Modifying a rule in one dimension causing multiple dimensions rules to be run (issue #3601).
  45. Transactional extract performance has been increased on certain types of extracts by up to a factor of 10 (issue #3603).
  46. Under some circumstances, if the viewer was terminating while a transactional extract was running, the engine would fail to release the dataset after the last user disconnected, assuming the dataset was not sticky (issue #3607).

Loader

  1. New feature: The loader user interface has been completely overhauled with a focus on usability and clarity (issue #3350). In some cases it may be necessary to use the legacy dataset loader tool to work around a bug or missing feature in the new loader. To launch the legacy loader from the viewer hold down the shift key when clicking on the Dataset menu in the upper left. This will show a menu option called 'Legacy dataset loader' which gives the option to open the legacy loader in either the main screen or to modify the current dataset.

Legacy loader/dataset knitter

  1. The loader will no longer create duplicate nodes when the input field name is more than 255 characters long (issue #3057).
  2. Thinning an index when no dependent dimensions or measures existing would cause transactional extract referencing the index fields to display blanks and N/A’s (issue #3522).
  3. In some cases, using single quotes and a non "-" char for a separator character when linking an index expressions causes an sql syntax error (issue #3535).
  4. In the advanced functions screen the right-click context menu was very slow to initialize for some datasets (issue #3540).
  5. Clicking the [Back to viewer] button would sometimes show the error message "Communications with the BIQ Serer were lost. (issue #3549)"
  6. Distloader leaving incorrect data in dimension.hiearchykey field when creating hier-node dimensions (issue #3552).
  7. Modifying a field type and then showing a data set report causes a hang (issue #3555).
  8. Changing to a different Hint format causes an error when the field type is already a date (issue #3556).
  9. It was possible to add a duplicate fact field name with different upper/lower case (issue #3578).
  10. During merge the secondary cube lost its view icons (issue #3584).
  11. The loader process could fault if the current user does not have write access to files created by another user (issue #3600).

Chopper

  1. New feature: The File Transform Utility (chopper) user interface has been overhauled with a focus on usability and clarity (issue #3517).

Release 5.0.9440 (changes since 5.0)

Viewer

  1. In rare circumstances, data blocks would not size correctly after drilling (Issue #3301).
  2. View descriptions are now multiple lines, RTF format and editable (Issue #3301).
  3. Selected changes to meta measures no longer cause the viewer to restart (Issue #3301)
  4. When connecting to engines with versions earlier than 4.05 running reports causes fault (Issue #3475).
  5. After exiting viewer book function with preexisting filter position, a fault would sometimes occur (Issue #3478).
  6. Saving a view name with single quotes in it causes a crash (Issue #3481)
  7. The viewer now handles request for data from hidden dimensions when updating requests, automatically (Issues #3484).
  8. Charts within Excel tables are now forced to update after the table is filled (Issue #3485).
  9. Importing extracts containing multidimensional tables cause extract items to be lost (Issue 3490).
  10. Changing views when reference sets were active caused data block sizing issues (Issue #3491).
  11. Closing all blocks and then exiting would cause earlier versions of the viewer to become confused during initialization (Issue #3496)
  12. Exporting extremely large datasets could cause a fault (Issue #3497).
  13. When starting a second viewer, and only one viewer was allowed, the second viewer would not completely exit after displaying the warning message (Issue #3499).
  14. Attempting to delete a filter set without first selecting one, would complain and then do nothing (Issue #3500).
  15. A “/” in a snapshot name would not allow the snapshot to be saved (Issue #3503).
  16. The restoring snapshot progressbar was not working property (Issue #3504).
  17. Scatter / gather would sometimes cause data block hidden status to become confused (Issue #3505).
  18. The pink mapping screen is always on top (Issue #3507).
  19. In rare circumstances snapshot restore would display an erroneous error message (Issue #3508).
  20. Double clicking on the viewer’s main form header did not maximize the window (Issue #3510).
  21. New option allows ignoring non-fatal errors when importing extracts (Issue #3514).
  22. Several dimension display options on the meta-measure form (f(x)), did not work correctly (Issue #3515).
  23. Defining transaction level field selection did not work correctly (Issue #3516).

Loader