Despite the fact that many regard Postgres as the best open source database platform I find myself frustrated by it’s lack of standard functions.  I understand that Postgres is designed to be extensible and that user defined functions can be built but I need my code to be both portable and read-only so I have to work with what I’m given.  Ideally what I’d be looking for is an equivalent of Microsoft SQL Server’s ISNUMERIC() or Excel’s ISNUMBER() functions but very much like MySQL I had to turn to regular expressions although as you’ll see, Postgres does not have a clean and clear REGEXP() function…

SELECT DISTINCT contact_number
FROM customers
WHERE (contact_number ~ ‘^[0-9]+$’)

I hope that helps any of you out there that encounter the same problem, thanks to the poster here for my original answer.

Despite the fact that the data looks like text whenever I tried to return the data into an application, I tried SSIS, SQL Server Linked Servers and Business Objects, each time the data would come back unusable or an error would be returned.  Business Objects gave me the key by declaring “This is a BLOB.” as you can see in the following screenshots…

SQL Server Integration Services

SQL Server Linked Server

Business Objects Desktop Intelligence

As with my recent post about loading data into MySQL with SSIS the saviour turns out to be an ODBC configuration setting, this time in the Metadata tab of the MySQL ODBC driver.  All you have to do is check the “Always handle binary function results as character data” and instantly your problems will be solved…

Out of a crazy fit of completeness I also took screenshots of the final results and it’d be a shame to waste them so here they are…

SQL Server Integration Services

SQL Server Linked Server

Business Objects Desktop Intelligence

Unfortunately I hit a slight snag on the first simple attempt, I could read the data without a hitch but despite multiple attempts at tweaking SSIS settings and reconfiguring connections I could not manage to get data into the MySQL target table.  Amongst other messages in the Output window in BIDS I could see the following error…

[MySQL][ODBC 5.1 Driver][mysqld-5.0.45-log]You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near ‘”tblDestination” (“col1″, “col2″, “col3″, “col4″, “col5″, “c’ at line 1

After some Googling I determined that the issue was down to an incompatibility between the SQL being generated by SSIS and the default MySQL engine, specifically that MySQL was not setup to accept ANSI compliant SQL.  Most of the solutions out there seemed to require that the entire server was configured in the ANSI compliant mode which was a problem for me as it’s a shared server and I can’t go reconfiguring it at will.  Thankfully MySQL provides the ability to set the mode on a per connection level with the following command…

SET SESSION sql_mode= 'ansi';

The trick is – how to get SSIS to run this before inserting the data without messing with the package itself?  The easiest way is to have the ODBC driver send the command at the point SSIS opens the connection, this can be configured in the settings of the ODBC source…

This is a relatively common feature of ODBC drivers and allows you to setup initial configuration options for this exact reason, now I know that I will not impact any other users of the server and my SSIS package works just fine.

Msg 7399, Level 16, State 1, Line 1
The OLE DB provider "MSDASQL" for linked server "PG_SERVER" reported an error. The provider reported an unexpected catastrophic failure.
Msg 7350, Level 16, State 2, Line 1
Cannot get the column information from OLE DB provider "MSDASQL" for linked server "PG_SERVER".

After some digging I came across a handy article on Microsoft Connect describing the same issue, with thanks to Nenea Nelu here’s the solution…

• Expand Server Objects > Linked Servers > Providers. 
  
• Right-click on MSDASQL and select Properties…
• In the Properties dialogue un-check “Allow inprocess” as follows…
  
• Click OK and re-run your query. 

Hopefully that should solve your problem, please note that this will affect all Linked Servers using that provider however as the Connect article points out – this is best practice for linked servers anyway.

I’ll follow up when I get the chance with a brief description of the process I followed but for now, please grab your phone and give it a try. Let me know if you have an issues, I only had a chance to test it briefly since I’m an iPhone user!

Download Veggie Phrases for Android

Anyway, to get to the point – today I stumbled across an intersting blog by Slawomir Chodnicki covering Open Source BI and ETL products from Pentaho, JasperSoft and SQL Power Group.  Much of the content is of the ‘how to’ variety but if you’re interested in the topic as a whole it’s worth stopping by Adventures With Open Source BI. 

This would be pretty straight-forward in Excel with the ISNUMBER() function, or in T-SQL with ISNUMERIC(), but neither work in MySQL so after a little searching around I came across this solution…

SELECT DISTINCT contact_number
FROM customers
WHERE contact_number REGEXP (‘[0-9]‘)

Effectively we’re processing a regular expression on the contents of the ‘contact_number’ field, it may seem like using a sledgehammer to crack a nut and I’ve no idea how performance would differ from a more simple approach but it worked and I guess that’s the point.

Market Trends

To start, I’ve taken a series of snapshots from the excellent Google Trends showing global search volumes for each of the Big Four offerings to measure the level of interest.  It’s reasonably clear to see from the graph below that interest in OBIEE shows a small but steady growth whilst Reporting Services shows a marked decline and the other two offerings remains roughly static (maybe a small decline?), this surprised me given that with the release of 2008 R2 I think that Reporting Services is really getting to the point where it offers a legitimate choice in the BI marketplace.  Perhaps the issue that Microsoft have fragmented their BI offering to include a mixture of terms with Excel, PowerPivot, SharePoint, Analysis Services and Reporting Services all making up the BI stack and nobody really knows what to call it?

This year has also brought an increased emphasis on Mobile BI with the iPad and iPhone fast becoming common executive playthings, Business Objects making it’s Explorer and Xcelsius products available on Android in addition to the iPhone (Explorer only).  MicroStrategy took the mobile emphasis a step further (perhaps to help stick their head above the crowd) by announcing a strong focus on the mobile BI market and offering a free 25-seat licence for their Mobile Suite.  Despite a strong focus on marketing Mobile BI I’m still not convinced that any of the vendors have really hit the nail on the head with their solutions in that whilst many offer pretty visualisations and slick interfaces most seem to lack the kind of simplicity that helps to present information quickly and succinctly, even the frankly beautiful independent product RoamBI just feels a little overdone when it comes to actually using it.

Major Product Releases

It’s been quite a year in the BI & Database world with the launch of Microsoft SQL Server 2008 R2, Oracle Business Intelligence Enterprise Edition (OBIEE) 11g and IBM’s Cognos 10…

Microsoft’s launch is effectively a moderate evolution of SQL Server 2008 in most areas with little change to the database engine, it’s ETL tool Integration Services and it’s OLAP engine Analysis Services.  That said, R2 did bring some handy incremental features which will be especially welcomed by the budget-conscious with an increase in the DB size of the free Express Edition from 4GB to 10GB and the addition of Backup Compression to Standard Edition.  There were some interesting additions with PowerPivot, Master Data Services and StreamInsight thought I’m not sure that either will find a natural home for a good year or so as busy DBAs and developers struggle to find the time to try these new features out.

Despite the major jump in the version number Oracle’s release too seems to be mainly an evolution and as a great fan of the product I’m quite considerably relieved since Oracle could quite easily have been over-zealous in integrating their ‘own’ tools like Discoverer and Warehouse Builder with bought-in technologies like Siebel Analytics (which became the bedrock of OBIEE), Hyperion’s Essbase and Sunopsis (now Oracle Data Integrator).  One of the less exciting but fundamentally important additions is that the semantic layer employed in OBIEE will be directly and immediately compatible with future releases of other Oracle products in the CRM, ERP and Finance application spaces.

I’m not as familiar with Cognos as the other two tools having only experimented with Cognos 8 for a couple of weeks but from everything I’ve read it seems that Cognos 10 was certainly a major milestone in the product’s lifecycle.  Aside from the shiny sounding features such as Social Networking and iPad support (actually a very serviceable looking mobile BI app) there are some very cutting-edge additions to the product including a statistical engine drawn from SPSS and Active Reports which allows users to explore and analyse offline data including interactive email reports.

The Future?

No good review and roundup article ends without a nod to the future and whilst I’m not keen on making absolute predictions there are a few developments I’ll be keeping my eye on for 2011 and beyond.

The main event I’m anticipating is the release of Business Objects XI Release 4, I’ve not seen too many concrete details about functionality but over the last few years Business Objects have seen themselves distracted by the Crystal acquisition (including the shoe-horning of their core product into Crystal Enterprise) and in turn their acquisition by SAP.  As a regular and long-term user of Business Objects I’m really hoping that they’ll blow away some of the cobwebs and deliver some new functionality as well as rounding off some of the edges that in previous versions feel a little unfinished, it would be great too if they finally included the key functionality from the legacy desktop client (which many long-term customer still rely on) in their core Web Intelligence product (Freehand-SQL & Grouping – I’m looking at you).

Another area to watch in Business Intelligence and Data Warehousing as well as the wider enterprise market is cloud computing, Informatica’s ETL in the Cloud offering has seen improvements and adoption throughout 2010 and it’s almost a given that Microsoft will be adding some degree of ETL capability to their SQL Azure platform. I’d expect an announcement if not a release along these lines in the coming year, though it’s possible that ETL comes behind providing cloud based analytics (something SSIS guru Jamie Thomson suggests).

In a broader sense I’m expecting to see a little more interest and pickup in the open source BI market, I’ve been saying this for a while (“this time next year, Rodders…“) and I might be wrong for some time to come but I always keep an eye on companies using an Open Source model such as the ETL vendor Talend who recently acquired Sopera (a middleware and SOA vendor), BI vendor Jaspersoft and all-rounder Pentaho.  With the global economy still suffering a hangover from the sub-prime mortgage crisis and banking collapse people have been looking for cheaper alternatives and open source companies provide a great way to achieve that, though some of Talend’s high-end offerings are almost comparable in price with other commercial products.

Another possible area to watch out for is the area of Personal Intelligence, essentially Business Intelligence for the individual.  A colleague and I spoke about this back in 2008 and we could both see that as people increasingly become data-aware they’ll start to look inwards and aim to measure things about themselves, one obvious starting point is fitness and we already have sites to log and chart your weight and calorie intake as well as the brilliant Nike+ product that measures your pace, time and distance during a run using either a sensor in your shoe or GPS (iPhone app), see the sidebar of this blog or below (one of my runs on the Nike+ site) for examples of the output.

I obtained the latest driver (5.1.7) from MySQL’s standard ODBC Connector page, installed it and added a System DSN without a hitch, the odd part came when Business Objects was returning only one row from a query that should return a couple of hundred.  Having run the same query on XP (I’m parallel running now) I suspected that the problem must be with the ODBC configuration so I attempted to delete the DSN only to receive the “Invalid attribute string” error…

A little Googling later led me to MySQL Bug #56233, in the discussion Fred Zappert frames the simplest solution suggested which is to uninstall version 5.1.7 and install version 5.1.6 instead.  Oddly MySQL don’t make it especially clear how to get hold of previous minor versions of the ODBC drivers but you can get it from here: mysql-connector-odbc-5.1.6-win32.msi 

Alternatively you could always visit bisql.net’s Tools, Utilities and ODBC Drivers page where I have a link for 5.1.6 which I’ll keep in place until the next Windows 7 compatible driver is released.

The second talk was Matt Whitfield‘s ‘CLR Demystified’ and not being a developer I’ll admit that much of it went over my head but I get the basic principles and most importantly I know what’s possible and where I might make use of CLR. The most interesting avenues for me are the ability to write custom aggregate functions and define custom data types – I was also impressed with the opportunity to increase performance in certain text processing / forward log parsing situations.

The next talk I attended was a gentle introduction to PowerShell and James Boother did a good job of showing where it might be useful for admin tasks, in particular the example of purging old backup/log files based on age lit my eyes up. There was also a demo of PowerShell authenticating with Twitter and posting tweets as admin alerts, alas the demo failed but that could well have been a timeout on the Uni’s WiFi network or just plain old demo-gremlins. I must say thou that I still can’t view PowerShell with some sense of disappointment because with all it’s flexibility and power it’s so damned wordy and many of the tasks it performs could be achieved with less code that a good old Bash script.

As with Friday I attended Quest‘s lunchtime session run by Kevin Kline, Ian Kick, Brent Ozar and Buck Woody – they’re some of the most experienced guys in the SQL community and when you get them together they’re funny as he’ll too so I was both entertained and informed in their myth-busting quiz.

My first afternoon talk was Gary Short’s session on NoSQL which predictably sparked a few polite but irate rebuttals from argumentative DBAs but the session itself was an excellent whistle-stop tour of the predominant NoSQL technologies and use cases. I was encouraged to hear from someone experienced in the field that nobody has quite put together all the pieces to hook up BI tools (that traditionally expect relational/dimensional models or OLAP sources) to the NoSQL back-ends, it’s a shame since I might need to do so pretty soon – I guess I’m going to have to get my hands dirty then!

The final talk was from Kevin Kline of Quest who covered SQL Injection, it was an informative talk that gave me pause for thought about a couple ‘best practices’ that I probably ought to harden a little. Kevin recommended a few tools that I’ll definitely be checking out at some point, notably:

• HP Scrawlr
• URLScan
• Source Code Analyser for SQL Injection
• Assessment and Planning Tool
• Discovery Wizard

And a few handy sites/articles:

• SQLSecurity.com
• MSDN ms161953
• MSDN ms998271
• ferruh.mavituna.com

Once more it was a great conference and the free day was every bit as good as the paid day, I can honestly say that I walked out of the event already looking forward to the next one.