Slicing overlapping subgroups of data with a synthetic “Filter Table”

I saw a post from Marco Russo (Twitter) showing a trick to use a slicer to filter across multiple columns at once in Power BI with some DAX and a new dynamic filter table. I’ve done something similar in the past to use a slicer not to slice data directly, but to change a view in the data or change filtering criteria indirectly – I wanted to share an example here to show a slightly different use case than Marco’s for a similar technique.

I want to create a slicer that switches the dataset between different subsets of data, but with overlap – like just my sales, sales for my group, and sales for the company. This is normally accomplished with a calculated column for the group the data is in, along with a slicer where if you want to see everybody, you have to select all three values like this:

But what if you want to create slicer options that aren’t exclusive? Where selecting “My Department” includes you too? Here’s how you do it!

In my example, I’ll need a fake sales and person table, so I’ll just use the “Enter Data” interface to create them:

Creating a sales table
Creating a user table (just a list from the sales table)

At this point, we could just join them on name and add some slicers, but here’s where we get more advanced by creating the filter table. The names repeat, but they show once for each slicer option they belong with, so whatever you select, you’ll see the relevant people:

Creating the filter table with all the slicer options in it

Once you have those three tables, you need to link them up – note the two-way relationship between the filter table and the person table. Without that, filtering on the last table won’t flow through to the sales table and our slicer won’t work:

Create the relationship and make sure the filter-user link to bi-directional

That’s all it takes – now you can create a slicer on the “Filter” column in the “Team Filter” table, and you’ll get whatever non-exclusive subset of the data you’re looking for!

I’m looking at only Bob’s sales
I’ve added Sally’s sales, but can still see Bob
Now I’m looking at everybody!

I made the slicer single-select in this example, which I do to avoid confusion – while this still works if you leave it a normal multi-select, it can lead to some confusion since users will expect the options to be exclusive and if they’re already got “Everybody” selected, neither of the other two options will do anything 🙂

If you want it, grab the PBIX file here.

I hope this is helpful to somebody – if you have any suggestions or other use cases for this, please add them below!

Choosing which version of Power BI Desktop opens a PBIX file

When you have both the “Report Server” and “Regular” version of Power BI Desktop installed, double-clicking on a PBIX file will open it in whichever one you updated last (most likely the non-RS version, since its updated every month). Unfortunately, I want to open PBIX files in my RS version, since that’s where most of the development for my job occurs.

I got tired of fixing the shortcut links every time I updated the Desktop, so I just added two new right-click options for PBIX files – “Start in RS version” and “Start in regular version. When it’s done, it looks like this (with the new options highlighted in red):

To add them, just take the code below and save it into a file with a .REG extension, and then double-click it to add it to your registry.

Windows Registry Editor Version 5.00

[HKEY_CLASSES_ROOT\PowerBI.File]
@="Microsoft Power BI Desktop Document"

[HKEY_CLASSES_ROOT\PowerBI.File\DefaultIcon]
@="\"C:\\Program Files\\Microsoft Power BI Desktop RS\\bin\\PBIDocument.ico\""

[HKEY_CLASSES_ROOT\PowerBI.File\shell]

[HKEY_CLASSES_ROOT\PowerBI.File\shell\Open in RS Version]
@="Open in RS Version"

[HKEY_CLASSES_ROOT\PowerBI.File\shell\Open in RS Version\command]
@="\"C:\\Program Files\\Microsoft Power BI Desktop RS\\bin\\PBIDesktop.exe\" \"%1\""

[HKEY_CLASSES_ROOT\PowerBI.File\shell\Open in regular version]
@="Open in regular version"

[HKEY_CLASSES_ROOT\PowerBI.File\shell\Open in regular version\command]
@="\"C:\\Program Files\\Microsoft Power BI Desktop\\bin\\PBIDesktop.exe\" \"%1\""

Let me know if you have any issues!

Import first sheet of an Excel workbook into Power Query, regardless of its name

I recently had to import a large number of Excel workbooks into Power Query (using “From Folder” – if you haven’t used this feature, it’s a huge time-saver when you’ve got a ton of matching files to import). The problem was that each workbook had a different name for the first tab and the import process gets data from tabs with the same name (assuming all your files match).

By default when you create a workbook, Excel names the first sheet “Sheet1”, but if the files you’re importing have sheets with different names, you can tell Power Query to load the first sheet of each file, regardless of the name – here’s the process (skip to step 5 to see the specific M-query solution):

  1. Select “Get Data” -> “From File” -> “From Folder” – this lets you import all the files from a folder at once (assuming their format matches).
  2. Select the folder where your files are stored (I have three files in the folder in this example).
  3. Select an example file and which sheet you’d like to import – this is what you’ll use to create your transformations (before the files are merged together) – in this example, my sheet is called “Bob’s sheet” (which is specific to just that single file).
  4. When you select “Import”, you’ll receive the error “[Expression.Error] The key didn’t match any rows in the table.” – this is because the other workbooks don’t contain any sheets with that name (though the error is a bit cryptic).
    If you double-click on the “Data Files” query with the error, you can see that it loaded the data from the first file, but nothing from the others:
  5. To correct this issue and tell Power Query to just use the first tab in the workbook in every case, open the “Transform Sample File from Data Files” query and click “Advanced Editor”.
  6. You’re now looking at the M-query behind this object (if you want it, here’s a great video on M if you’ve never worked it before and want a primer – I watched a bunch of different videos and this one really connected the dots for me). The section with the sheet name is highlighted:
    Replace the highlighted text with “Item=Source{0}[Item]” so that it looks like this:
    (The “0” refers to the first sheet – if you want the second sheet, you’d use “1” and so on)
  7. Once you make change, you can see in the Power Query window that it’s now getting data from every file.
  8. Once you click “Close and Load” to get back to Excel, click on the “Refresh All” button to force a data refresh and you’ll see it import all the files, regardless of their sheet name:

That’s it – you’re pulling in all the workbooks regardless of their sheet name!

I’d never used the “From Folder” feature, but I’d definitely recommend trying it out – it was a huge time-saver for me and allowed me to import hundreds of matching Excel files into a single model in minutes!

Client certificate error with Office 365 and Blackberry Hub

Our office recently implemented two-factor authentication with Office 365 and it caused some issues with my Activesync account in Blackberry Hub (running on my Android Pixel 2. I would receive email intermittently – an hour on, a few hours off, and then a random bunch of email again. Then I started receiving the following error message:

The client certificate for (your mail account) must be allowed for use in establishing secure email connections. Do you want to do this now? [Password may or may not be required.]

Before the two-factor change, out email configuration required that we user a client certificate that was issued to our phone through Microsoft Intune (called “Company Portal” on our phones). Once two-factor was enabled, leaving the certificate selected gave that error message. Instead, email needed to have the certificate removed as follows:

Here was the big change: instead of my domain password, Hub required me to generate an app-specific password using the following process:

https://support.office.com/en-us/article/create-an-app-password-for-office-365-3e7c860f-bda4-4441-a618-b53953ee1183

The mobile version of office.com doesn’t allow you to generate an app password, but you can tell the Chrome on your phone to view the full desktop version of the site and it will allow you to generate a password. Copy it, use it in Blackberry Hub (without the certificate selected), and you’re back to receiving your email!

Passing multi-value report parameters to SSRS using the URL querystring

Passing parameters to reports in SSRS using the URL can be really helpful, especially when you want to link to a report with some of the data aleady filled out. We had an email alert that some accounts in active directory were incorrect, and we wanted a way to link to a report that provided some additional detail for those accounts. Querystring parameters aren’t very well documented (and I couldn’t find anything on multi-value parameters), and I hope this saves you some time if you’re in the same situation.

There’s a pretty good reference on Technet, but it’s got some gaps (for me, the biggest was even mentioning multi-value/multi-select parameters). The basic syntax is pretty straightforward – normally, a report is called with a URL like the following:

http://YourSsrsServer/Reports/Some%20Folder/YourReportName

From there, you can select the parameters and run the report. To provide a link to the report with the parameters already filled out (if they’re blank or if you gnat to override the defaults), just append this to the end of the URL:

?ReportParam1=SomeValue&ReportParam2=SomeOtherValue

You can pass any type of parameter with this same syntax – text, datetime, integer, float, or boolean (use Yes/No). Your full url would look like this:

http://YourSsrsServer/Reports/Some%20Folder/YourReportName?ReportParam1=SomeValue&ReportParam2=SomeOtherValue

To do a multi-value parameter, you repeat the name of the parametr to provide it multiple values:

http://YourSsrsServer/Reports/Some%20Folder/YourReportName?ReportMultiParam=SomeValue&ReportMultiParam=SomeOtherValue

If this is the only parameter on the report and it doesn’t need any other user input, the report will automatically execute because it has everything it needs. If you want the report to wait for the user to hit “View Report”, you need to leave at least one parameter empty.

Two quick notes if you’re having trouble:

  • Use the parameter name, not the prompt text. If your parameter is called “Office” but the prompt says “Which office?”, use “Office”. Parameter names aren’t always intuitive and you may need to open the report in Visual Studio to get these names if you can’t guess them.
  • In the same way, you need to use the field’s value, not the label. If the dropdown contains the full name of the office, but the value behind the list is the office number, you’ll need to provide the numbers on the querystring.

Good luck!

Microsoft Connect item for an SSRS pre-report canvas

Current state:

About six months ago, I submitted a connect request for SQL Server (my first one) to suggest an improvement to the SSRS interface – a customizable canvas that’s displayed before an SSRS report renders, while parameters are being entered. My issue was that there was information I wanted report users to have (like expected wait time, suggestions for the parameters, the report title and a short description that might help them, etc.) and there was no way to get it to them.

With a giant empty screen staring at them, it seemed like a great place to add that information:

SSRS Report pane current layout

The suggestion:

Ideally, while the user is considering parameters, I’d love to see something like this (ignoring that it’s not a super-photogenic report):

proposed-layout

This would be a great place for details about the report that help the user before (or while) they run it, like:

  • The name of the report (not always clearly visible in the URL or browser title bar, especially when you’re Sharepoint-integrated. Especially helpful if they’ve left multiple report tabs open or if report generation fails, they’re often left with a white screen and no easy way to tell the tabs apart.
  • A short description of the report, including when you’d use it or any notable caveats.
  • An estimated runtime – this isn’t always possible exactly as parameters vary, but even a general estimate (or better yet, an average or a 95% confidence range, pulled from the report server database) would be better than the complete blank they get today.
  • Suggestions/restrictions on the parameters – If a report can’t be run for all offices at once, you can say so here. If running it for everybody at once means you can’t export to Excel because it’s too large, let the user know.
  • A mock-up/screenshot of the report – so they user can see if it’s the one they want. They may be looking for a report with a specific chart, and waiting for 3 minutes for each one to render may not be ideal.
  • A link to any other documentation/resources, or related reports, or pretty much anything else.

It wouldn’t need to be anything fancy – I’d even settle for text you can format, or some kind of additional panel on the report that you can design, but isn’t rendered with the report itself. Access to data (to pull estimated on execution times or dynamic owner/description details, for example) would be nice, but not required – I’d rather have a design-only, no-data canvas than nothing at all.

The plea:

t-sql-tuesdayThanks to the motivation from the folks at Brent Ozar Unlimited, I’m asking for your help in voting for this suggestion (if you think it would be beneficial, of course). As part of T-SQL Tuesday, they’ve asked people to find a connect request that they’d like to see filled and post about it, so here I am asking humbly for your support. : )

If you like the idea, please vote for it – if you don’t, please comment and let me know why not (I’m always open to understanding the opposition). There are a ton of great Connect items out there and this is only one, but I think it would help out (especially for our longer-running reports).

Thanks for your help!

Direct URL for the Connect item: https://connect.microsoft.com/SQLServer/feedback/details/2809098/ability-to-customize-report-splash-screen-during-parameter-submission

Side note:

I’ve attached the two images (current/proposed) to my connect item three different times and they’re still not showing up there – if anybody knows how to get those images published, people can see what I’m talking about when I describe my suggestion! : )

Query Business Objects repository for list of reports

We’re migrating away from Business Objects Enterprise (XI-R2, not the newer 3.X version) and into SQL Reporting Services (it’s about time!) and I needed an inventory of everything that was still in BOE. The repository is obfuscated so that it’s not human readable, but thanks for some sleuthing and an 8-year-old post from a Business Objects forum, I wanted to share the script I was able to come up with.

The hard part (decoding the obfuscated text from the database into human-readable format) came from a post by “marklee” on a Business Objects board (his post is about 1/4 of the way down):

http://www.forumtopics.com/busobj/viewtopic.php?t=124272

His query was also written for Oracle, so I converted it to one that would run against SQL Server, and then I added some additional content types based on what we had in our database, and then attached a few additional rollups (like showing the root folder, building out the full path to the report, and returning the number of instances of each report that had been generated.

You can download the full script if you’d like, or you can view the contents below. Enjoy!

select 
 objectid, 
 parentid, 
 ownerid, 
 convert(datetime, REPLACE(left(lastmodifytime, 10),' ','-') + ' ' +
			replace(substring(lastmodifytime, 12, 8), ' ', ':') +
			'.' + substring(lastmodifytime, 21, 3)) as lastmodifytime, 
 upper(left(objectname,len(objectname)-2)) ObjectName, 
 typeid, 
 type, 
 convert(bit, 0) as IsInstance,
 si_cuid 
 into #results
 from 
 ( 
 SELECT 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 replace( 
 cast(objname as varchar(2000)) 
 ,'S','v') 
 ,'M','s') 
 ,'A','m') 
 ,')','a') 
 ,'+','b') 
 ,'C','n') 
 ,'-','c') 
 ,'/','d') 
 ,'O','t') 
 ,'E','o') 
 ,'1','e') 
 ,'3','f') 
 ,'G','p') 
 ,'5','g') 
 ,'7','h') 
 ,'W','x') 
 ,'U','w') 
 ,'Q','u') 
 ,'I','q') 
 ,'9','i') 
 ,':','i') 
 ,';','j') 
 ,'K','r') 
 ,'=','k') 
 ,'?','l') 
 ,'[','y') 
 ,']','z') 
 ,'!@',' ') 
 ,'B~S','&') 
 ,'!BO','.') 
 ,'B|C"','(') 
 ,'!B|D',')') 
 ,'M|Z','-') 
 ,'M}L',',') 
 ,'M|N','_') 
 ,'M}Z',':') 
 ,'!B{B','''') 
 ,'`|<','0') 
 ,'`|>','1') 
 ,'`|@','2') 
 ,'`|B','3') 
 ,'`|D','4') 
 ,'`|F','5') 
 ,'`|H','6') 
 ,'`|J','7') 
 ,'`|L','8') 
 ,'`|N','9') 
 ,'{','') 
 ,'!','') 
 ,'"','') 
 ,'@','') 
 ObjectName, 
 convert(varchar(100), case 
	 when TypeID = 1 then 'Folder' 

	 when TypeID = 2 then 'Crystal Report'
	 when TypeID = 260 then 'Web Intelligence Document'
	 when TypeID = 290 then 'Object Package' 
	 when TypeID = 265 then 'Report Instance (TXT)'
	 when TypeID = 262 then 'Webi Report' 
	 when TypeID = 314 then 'Deski Report' 

	 when TypeID = 283 then 'PDF' 
	 when TypeID = 267 then 'Text' 
	 when TypeID = 323 then 'Excel' 
	 when TypeID = 319 then 'Excel (XLS)'
	 when TypeID = 259 then 'Word (DOC)'
	 when TypeID = 279 then 'Report Instance (PDF)'

	 when TypeID = 266 then 'Universe' 
	 when TypeID = 278 then 'Publication' 
	 when TypeID = 299 then 'Connection' 
	 when TypeID = 8  then 'Shortcut' 
	 when TypeID = 20 then 'Groups' 
	 when TypeID = 13 then 'Server' 
	 when TypeID = 16 then 'BO Server' 
	 when TypeID = 21 then 'Event' 
	 when TypeID = 24 then 'License Key' 

	 when TypeID = 19 then 'User type 19' 
	 when TypeID = 18 then 'User type 18' 
	 when TypeID = 47 then 'User type 47' 
	 when TypeID = 48 then 'User type 48' 
 
	 else 'Other' 
 end) Type, 
 * 
 FROM [dbo].[CMS_InfoObjects5] 
 where typeid <> 284 -- appears to be some kind of temporary file
 ) BORepository 

create clustered index #idx_c_results on #results (objectid)
create index #idx_results_parentid on #results (parentid)
create index #idx_results_ownerid on #results (ownerid)

-- Mark the ones that are instances of another object (children of non-folder)
update r
   set r.IsInstance = 1
 from #results r
join #results r2
  on r.ParentID = r2.ObjectID
where r2.TypeID <> 1
  and r.TypeID <> 1




-- Define all the levels of the objects and add them to a lookup table
DECLARE @ObjectLevel INT = 0

select 23 as objectid, 0 as parentid, 0 as ownerid, GETDATE() as lastmodifytime,
		convert(varchar(100), '$') as objectname, convert(int, null) as ChildInstances, convert(datetime, null) as LastInstanceTime,
		1 as typeid, convert(varchar(100), 'Root Folder') as type, @ObjectLevel as objectlevel 
into #objects

while @@rowcount > 0
begin
	;with p as (select parentid, count(*) as instances, max(lastmodifytime) as LastInstanceTime from #results group by parentid)
	update #objects
	   set ChildInstances = p.instances,
		   LastInstanceTime = p.LastInstanceTime
	  from #objects o
	  join p
	    on o.objectid = p.ParentID
	 where o.ChildInstances is null
	   and o.typeid <> 1
	
	SET @ObjectLevel = @ObjectLevel + 1

	insert into #objects
	select r.objectid, r.parentid, r.ownerid, r.lastmodifytime, r.objectname, null, null, r.typeid, r.type, @ObjectLevel as objectlevel
	from #results r
	join #objects o
      on r.parentid = o.objectid
	 and o.typeid = 1
	left
	join #objects o2
	  on r.objectid = o2.objectid
   where o2.objectid is null

end


-- Set up the object paths
select objectid, parentid, lastmodifytime, convert(varchar(4000), '') as ObjectPath,
		objectname, childinstances, lastinstancetime, typeid, type
  into #objectswithpath
  from #objects
 where typeid <> 1

while @@rowcount > 0
BEGIN
	update #objectswithpath
	   set parentid = o.parentid,
			objectpath = o.objectname + '\' + objectpath
	  from #objectswithpath owp
	  join #objects o
	    on owp.parentid = o.objectid
END

select objectpath, objectname, lastmodifytime, ChildInstances, LastInstanceTime, type
 from #objectswithpath
order by ObjectPath, objectname


--select * from #results
--select typeid, type, count(*) from #objects group by typeid, type order by count(*) DESC

/*
select type, case when isnull(ChildInstances,0) = 0 then 0 else 1 end, count(*), max(LastInstanceTime) --objectpath, objectname, lastmodifytime, ChildInstances, LastInstanceTime, type
 from #objectswithpath
 where ObjectPath not like '$\REPORTS OFF LINE\%'
group by type, case when isnull(ChildInstances,0) = 0 then 0 else 1 end
order by type, case when isnull(ChildInstances,0) = 0 then 0 else 1 end
*/

--select * from #objectswithpath where type = 'Word (DOC)'


drop table #objectswithpath
drop table #objects
drop table #results

SQL Agent job fails with QUOTED_INDENTIFIER error

If you have a SQL Agent job that’s failing, but the query succeeds when you run it manually in SQL Management Studio, it’s possibly a difference between the SQL Server default connection settings and those SSMS uses as the default. In my case, I was inserting from a stored procedure and received the following error:

[SQLSTATE 42000] (Error 1934) INSERT failed because the following SET options have incorrect settings: ‘QUOTED_IDENTIFIER’. Verify that SET options are correct for use with indexed views and/or indexes on computed columns and/or filtered indexes and/or query notifications and/or XML data type methods and/or spatial index operations.

To view the default connection options for your SQL Server instance, right-click on your instance in SSMS and select “Properties”, and then browse to “Connections” – you’ll see something like the following list (on my server, none of these are enabled by default for new connections):

Server Defaults

To compare it to the default settings for your installation of SSMS, click on “Tools” and then “Options…”, and then browse to “Query Execution” -> “SQL Server” and view the settings under both “Advanced” and ANSI” (In my case, SSMS was enabling a number of settings by default for my connections that SQL Agent jobs wouldn’t have enabled):

SSMS Results ANSI

SSMS Results Advanced

In my case, I just added an explicit “SET QUOTED_IDENTIFIER ON” to the beginning of the script in my SQL Agent job and it resolved the error message. Hope this helps you too!

Viewing Supplemental Logging status in your Oracle database

Viewing the status of supplemental logging on all of our Oracle tables is done via the all_log_groups table, but it’s not in an easy-to-use format and needs to be pivoted (and it only shows you which tables have logging of some kind, which makes it difficult to tell which tables don’t have anything). I’ve done that below to provide an easier view of the status of supplemental logging on all tables, regardless of their current status:

  select t.owner, t.table_name,
         max(case when lg.log_group_type = 'USER LOG GROUP'      then lg.always end ) as "User Log Group",
         max(case when lg.log_group_type = 'ALL COLUMN LOGGING'  then lg.always end ) as "All Column",
         max(case when lg.log_group_type = 'PRIMARY KEY LOGGING' then lg.always end ) as "Primary Key",
         max(case when lg.log_group_type = 'UNIQUE KEY LOGGING'  then lg.always end ) as "Unique Key",
         max(case when lg.log_group_type = 'FOREIGN KEY LOGGING' then lg.always end ) as "Foreign Key"
    from dba_tables t
    left
    join all_log_groups lg
      on t.owner = lg.owner
     and t.table_name = lg.table_name
   where t.owner = 'Table Owner'
   group by t.owner, t.table_name
   order by t.table_name

In the recent spat of Goldengate troubleshooting, I’ve had to check these values a number of times and I hope this query is useful to others (since I couldn’t find a good example online that made this information easy to view).

Please let me know below if you have any issues!

Correcting OGG-01733 in Goldengate – Trail file header file size mismatch

Warning – Oracle/Goldengate support will probably get mad at you if you try this. It worked great for me, but they recommended we reload the data from scratch, so that’s probably what they’d recommend for you too. Just know that this is 100% unofficial and unsupported 🙂

Quick summary:

We have Goldengate replication from Oracle 11.2 on Lunux to MSSQL 2012 on Windows, and we ran into an OGG-01733 error “Trail file header file size value {X} for trail file {Y} differs from actual size of the file ({Z})”, which caused an ABEND where we were stuck. We opened a ticket with Oracle support and after a week with very little response, the concluded that I should just perform a new initial load on the destination – since the trail files had already been pumped to the destination server and removed from the extract server, they were unable to troubleshoot further.

It turns out the work-around was to open the trail file in a hex editor and manually update the trail file header to make it think it was supposed to be the size it actually was. After saving the file again and resuming replication, it continued on its merry way and applied the transactions without another complaint.

Steps to resolve this error message:

  1. Make a backup of your trail file – you know, since you’re editing it and might want a second shot.
  2. Open the report file and make a note of the size the file is currently (“Z”) and the size it’s supposed to be (“X”). I’ll refer to those as X and Z further down.
  3. Use a decimal-to-hex converter like this one to convert both of these values to their hex equivalent (now I’ll call them “HX” and “HZ”)
  4. Load up the trail file in your favorite hex editing tool – I like using Notepad++ in combination with the HEX-editor plug-in (once the file is loaded, select “HEX-Editor” from the plug-ins menu, and then select “View in Hex”)
  5. Perform a search (if you’re using Notepad++, ensure the data type is set to “Hexadecimal”) for your “HX” value – the size the file thinks it should be. However, you need to search for an even number of digits – if your hex value is an odd number of digits, either drop the leftmost (largest) one or add a zero to the left (I dropped a digit):
  6. Goldengate - Hex Editor

    • Side note: You can see that my trail file size isn’t too far into the file – under 300 bytes from the beginning. However, since it’s stored in hex, it’s not something that’s easily viewable in the file (where you will see some file path and server version information if you look to the right where the ASCII is displayed. Also, in my image, the file size is preceded by a quite a few zeroes – my trail files are set to 100MB, but it appears Goldengate supports up to 4GB trail files using the 32 bytes in the header file. Back to fixing this…
  7. CAREFULLY edit the HX value you’ve found to be the new HZ value – the actual size of the file. In particular, don’t move any of the bytes around or add/remove anything, just fix the values you need to change so that the file size is stored in the same location.
  8. Save the file and close it.
  9. Resume replication right where you left off (assuming you made a backup and the edited the original trail file) – it should check the new file size, see the transaction that was previously beyond the file size limit, and then apply it and move on!

Conclusion?

What causes this behavior? I can’t find any clear documentation or explanation at all – when searching for this error, the only meaningful links I can find at all are either in an Oriental language and have basic details as well as a dire warning to call Oracle support immediately or a case where somebody receives it on an initial load and the forum’s advice is “your table is too small to mess with this – just export it to CSV and reload it that way”.

When we looked at the list of trail files, we noticed something particularly odd – the trail files near the offending file all had ascending “last modified” timestamps, as you’d expect, but this file was actually out of order:

05/01/2015  03:45 AM        99,999,462 SV002351
05/01/2015  04:38 AM        99,999,802 SV002352
05/01/2015  08:13 AM        99,999,367 SV002353
05/01/2015  10:09 AM        99,999,936 SV002354
05/01/2015  11:05 AM        99,999,630 SV002355
                                                 <-- File should be right here
05/01/2015  11:41 AM               891 SV002357
05/01/2015  11:47 AM        99,999,462 SV002358
05/01/2015  11:50 AM        99,999,280 SV002359
05/01/2015  11:58 AM        99,999,314 SV002360
05/01/2015  12:09 PM        99,999,910 SV002361
05/01/2015  12:40 PM        99,998,043 SV002362
05/01/2015  01:16 PM        99,999,754 SV002363
05/01/2015  01:34 PM        72,017,446 SV002356  <-- But it's down here
05/01/2015  02:05 PM        99,999,516 SV002364
05/01/2015  02:40 PM        99,999,966 SV002365

The file contained two additional transactions beyond the stated header size and the actual end of the file, and they were both time-stamped correctly to have been located in that file (they were both stamped 10:34AM, along with the transactions that were earlier in the file, and since the server is an hour off because of time zone, they were in the right file).

The fact that it’s smaller than the others, and that it’s followed by a file containing no transactions (just a header) led me to believe the file was cut short by a network interruption of some kind. We’re using a local extract and a separate pump, as we’re advised to do, but the connection still drops from time to time. In this case, I can only imagine it was in the middle of committing something, was interrupted, and then somehow these transactions were suspended for some reason and then added to the file later. I can’t imagine why, but when they’re added, the file header isn’t updated.

Hopefully this explanation and work-around have helped somebody else – we pulled our hair out for a week going back and forth with Oracle support and scouring the internet (unsuccessfully) for any relevant information – in the end, going rogue and editing the file was the only way (short of a complete reload) to get things moving again!