Added a new build and reporting engine based on JDBC and PostgreSQL. Contact firstname.lastname@example.org for assistance with switching to the new engine after upgrade.
56019 - A version of Adobe Acrobat was unable to open MITS Discover dashboard pdfs. Adobe has fixed the issue.
55762 - ABSOLUTE eons (those tied to a specific static time range, rather than relative to the build date) are not supported in JDBC mode.
56082 - Also in both modes when modifying a cube now there's also a button to go to its history.
55852 - Calculations that divides by zero is shown as blank. In the PICK engine, when that blank result is used in a secondary calculation, sometimes the result is blank and sometimes it's 0 depending on whether the column expression used a math operator like * or a function like MULTIPLY. In the JDBC engine, the secondary calculation will be blank.
55862 - Getting rid of spurious parsing errors for columns that use custom functions with column arguments.
56061 - If you switch to JDBC mode, build your cubes, and switch back to PICK mode, you'll need to rebuild them.
56045 - In both modes, the report listing page no longer lets you click grid reports or charts whose cubes are not usable.
56033 - In both modes, the support console includes a link for "Check JDBC Engine compatibility" which runs a tool that warns if you have configuration or content that would not be supported if you switched to the JDBC reporting engine.
56032 - In JDBC mode, hitting the chart button from a report may cause it to add FUSE to the drill down path in order to make there be one lowest level chunk of rows.
56055 - In JDBC mode, identifier values can be "fused" to merge row values together.
56051 - In JDBC mode, rows from external cubes will just show up on your reports without the need for MITS.MERGE.EXTERNAL.LINKS in the build process (which is not supported in JDBC mode).
56065 - In JDBC mode, trying to FLASH or build cubes at PICK results in 'The DISABLE.PICK.CUBE.BUILDING flag is set, indicating the gateway is not using the PICK reporting engine. Cubes cannot be built or flashed at TCL when this is true.'.
56056 - In JDBC mode, users with custom (CALL) restrictions can't log in, and are marked with warnings in user management.
56006 - In JDBC mode, when the lowest level is filtered out and bottom totals are on, you still see a total row even though it has no totals in it. In PICK mode it hides total rows when lowest level rows have been filtered out.
55737 - JDBC mode will log build steps, with optional debug logging available.
55838 - NOGRAND will show identifier-related columns (bang fields) at the hidden levels in JDBC mode.
55736 - Only JDBC build jobs are available in JDBC mode . Existing jobs created for PICK mode are shown with a warning icon.
56113 - Reduced verbose logging when exporting PDFs and sync-ing metadata cache.
55794 - Underlying logic for the connection between Java and PICK nas been modified. It should have no effect on behavior.
55918 - Using the JDBC engine, cubes stay reportable while they are being built. When viewing a Report, if you do anything that reloads data, if there is a new version of the cube, that data will be shown.
55909 - Using the JDBC engine, if a build fails, the last successful build remains reportable.
56037 - Using the JDBC engine, if a cell on a report has no transactions, we now display it as a blank instead of a zero. This permits distinguishing between, e.g., a customer that had no sales in the previous month vs a customer that was sold something at 0 dollars.
56107 - When creating a new build job from a paragraph, if the first task is a cube build, the cube's name is the default name of the build job. (Already worked for other build jobs, just also doing it for TCL paragraphs.)
51096 - When Omit Totals (NOGRAND) is in use along with hiding Zero Rows (NOZEROS) then rows are sometimes hidden incorrectly if they use use a Percent of Total column. This remains broken in the PICK engine but is fixed with the JDBC engine.
56007 - When selecting specific identifier values and specifying a value that does not exist, the engines behave differently. PICK mode shows zeros and JDBC mode shows blanks.
51207 - When using "No-data rows" showing (NOZEROS in the MQL), values that sum to zero can still have child rows. The only rows removed will be those with no backing transactional data underneath.
56012 - When using the JDBC reporting engine, the option to hide "zero rows" is replaced with an option to hide "no-data rows". This is similar but better, hiding all rows with no underlying transaction data (including rows that only show bang fields), but showing rows where transactions sum to zero.
55850 - With the JDBC engine, identifier-related columns (bang fields) do not affect which rows appear when no-data rows are hidden (NOZEROS in the MQL).
56095 - Updated PostgreSQL JDBC Driver to 42.2.8
The following are the operating systems, browsers, and databased we used to test this version of MITS.
Safari 12.1.1 (13802.2.6.1.2)
iPhone Safari 12.3.1 (basic read-only testing)
iPad Safari 12.4 (basic read-only testing)
Nexus Android 7.1.1 Chrome 74.0.3729
Windows 2012 Server
Windows 2016 Server
UV 11.1 UV 11.2.5 (with the patches) UV 11.3.1 UV 11.1.3 UDT 8.2.1 UDT 7.3.7
PostgreSQL for MITS Report:
For JDBC Mode: PostgreSQL V12 is required (although it tested fine on 10.5 and 11) For MITS Report, the following PostgreSQL versions were tested: 10.5, 11. 12