We Keep Coding

Index Design | Performance of Multiple Indexes VS 1 Index with multiple Data Sources

Implementing Azure Search. We currently got two tables (Relational DB) related to each other that we would like to use as an Index.
What's the best way to implement this?
Hitting Two indexes (One for each table), do the join between the
results in code.
Two Data Sources (One for each table), Two Indexers pointing at the same Index.
Using a SQL View, doing the join in it.
I'm using the first option as an Example, but I'm getting results between 200 to 400 ms for only 3 records -Using Postman. The business limit will be around 20 results. My concern with this is when I start to add more indexes and then do the join between them in the code could be not performant.
Say for example the main table, with 4 or 5 tables related to it (PK - FK relationship), I would have to create an index for each one of them. Is that OK? Would that cause performance issues? Should I try the second or third option?
Could you please expand your answer.
Thanks

I'd suggest you look at the new complex type capabilities of Azure Search. For 1:N relationships, you could put the related records in a Collection(Edm.ComplexType) field on the index. More info on complex types: https://learn.microsoft.com/en-us/azure/search/search-howto-complex-data-types
As far as indexers go, I'd recommend you create a view in SQL that pulls the related records for the current row into a JSON formatted text field.  You would then point the data source/indexer at this view and map the fields from the JSON field into a complex collection field in the index.
SQL actually makes defining the view pretty easy.  For example, for the Azure Search Hotels sample schema it would be: 
SELECT *, (SELECT *
    FROM dbo.Rooms$
    WHERE dbo.Rooms$.HotelID = dbo.Hotels$.HotelID FOR JSON AUTO) AS Rooms
FROM dbo.Hotels$
Hope this helps.
Mike Carter
Azure Search product team

SQL Server recommendation for storing segmented gl account codes in database

I've been tasked with an enhancement to our order system that will require importing segmented GL account codes for assignment on individual line items of an order.
I need to support querying the codes by segment1, segment2, etc in order to load cascading dropdown boxes for assignment by the user. The GL codes will have one or more segments delimited by a character. An example of a code is "1010.1034001.99.01".
I've loaded several thousand codes into a table for testing where the entire string value exists in one column (delimited by a character). I've created two variations of functions that return rows where segment1 value is equal to some parameter. The query also supports further querying by providing additional parameters for other segment values.
I intend to support these queries from the table using Entify Framework 6, but used sql functions to get a feel for what the performance may be when the GL account codes are stored in one column. Performance was not as good as I had hoped.
Does anyone have recommendations on how best to store this data (there may be 200,000 codes). Do you feel that I can query using EF and expect performant results?
Would a hierarchy organization make more sense for this data? Our team was hopeful to store the delimited values on one column.
Thanks in advance.

If you would use a table with three columns you could store the values cascading, enabling you to make your queries a lot easier and probably faster. Why would your team hope to store it in one column, what advantage does that have?
if you have
ID
Code
ParentCodeId
where ID is a unique key and ParentCodeId is a nullable reference to that unique Id you can split your exaple code as follows:
ID Code Parent
1 1010 null
2 1034001 1
3 99 2
4 01 3
By applying some logic when importing your codes, you can check if a code already exists as a parent on the needed level so you don;t have to repeat them, and that way you coul dget all codes that start with 10100 by selecting on selectiong on parentID 1.

EF and Linq with self referential table

I have a self-referential table in my database that looks sort of like above. Basically its setup in such a way that each row has a unique ID (identity PK) and a DependentID to indicate any other record in the set that it is dependent on. It is very similar to the parent-child type examples you often see in SQL textbooks but my case is subtly unique in the sense that a given record can also be dependent upon itself (see row 1 above)
Two questions:
Can EF be made to represent this relationship properly? I've read several posts on here that suggest that it does not deal with this scenario gracefully so my initial thought was that it might not even be worth it, I might be better off just treating it as a normal table and writing the business logic to ensure the data gets inserted/updated correctly. In my scenario, I won't ever be querying these entities thru EF really, the app will basically load them all at startup and then I'll run linq queries against them at runtime to filter as needed
Assuming I cannot get it to work with EF and as I note in #1 I simply load em all up into memory at startup (there are only going to be 50-100 or so), what would be the most efficient way to join on this via linq? I would want to be able to pass in a DependentId and get all the records associated with it and their properties...so in this example I'd want to pass in '1' and get back:
1 - John - 10
2 - Mike - 25
3 - Bob - 5
thanks for the help

Indeed, the entity framework cannot represent such a relationship, certainly not in in a recursively queryable form.
But you are not asking for recursive queries, so you could treat DependentId as just another data column. Doing that, it would be trivial to build and execute your question-two query against the database.
UPDATE:
That query would look something like
int dependentIdToSearch = 1;
var q = from something in db.mytable
where something.DependentId == dependentIdToSearch
select new { something.Id, something.Name, something.Value };
END UPDATE
If you do need recursive queries (all direct and indirect dependencies of), you need a table valued function with a common table expression. The entity framework cannot deal with that either, at least not in the current version. If you need this support, you can wait for EF 5 or use Linq to SQL (which had support for table valued functions since the first version years ago).
You can indeed also read the entire table in memory, provided that it is read-only, or that there is only "one memory" (single server, not load-balanced or client app with local database).
If it's read-only, you have the option to build an object graph once at load time, enabling efficient execution later. For example, you could define a class with a collection of objects that are dependent on each object. Your query then becomes a trivial iteration over that collection.

High performance Custom user fields

looking for examples/tutorial for custom user fields, not via EAV
EAV is going to be problematic for various reasons such as performance
there are many base entities/tables with over 100000 records each
there will likely be over a dozen attributes
the records are to be displayed in a flat ui grid incl. custom fields so flattening them would be an issue while maintaining performance
Looking at enabling this via DDL where all custom fields would go into a matching table such as
<tablename>_custom_<userid>
and all user attributes would map to a column each and all their metadata stored in a metadata table
the retrieval would be simpler where the query would simply be
select *
from <tablename> A, tableName_custom_userid B
where B.KeyField = A.KeyField --( perhaps using outer join, haven't gone that far yet )
Wondering if there are any gotchas down the road that i need to be aware of ?
of course any samples/pointers would be helpful to kickstart the effort
specifically would appreciate any advice on using DDL for Sql Server compact 4

One technique I have seen used is to use a sort of 'hard-coded' EAV pattern. Don't hang up! It worked well with the dataset sizes you were talking about and didn't actually use EAV - it was only EAV-esque.
The idea is to have a set of tables to store these custom attributes within it, with some triggers (described below) on them. The custom attributes tablesets store metadata about the attribute (what table it goes with, data type, constraints, etc). You can get very fancy with this but I did not haev the need.
The triggers on your meta-tables are there to re-generate views that rollup base+extension into first class objects within the DB. So instead of table person + employee extension table, you have an employee view that includes both. When you drop a new value into the custom attributes tables, the triggers will re-roll the views and include the new stuff. If you wanted to go nuts, you could also have the triggers re-write stored procedures as well. Depending on how your mid-tier code is structured, you would still be forced to re-code some, however this would be the case anyway should you be applying rules that read the data.
In testing, I found that for the relatively small # of records you're talking about, performance was somewhat slower but followed roughly the same pattern of degradation (2x the number of records, ~2x as slow).
-- edits --
How I saw it done, you had a table that represented your first class objects, so a row for 'person' and a row for 'employee,' etc. We'll call that FCO. Then you had a secondary table that stored what tables represented the FCO. We'll call that Srcs.. For person, there would be one row, which is the person table. For Employee, there would be two rows, the person table and the Employee extension. There is a third table, called Attribs, which stores the columns from the tables that constitute the FCO. For simplicity, we'll say Employee has ID, Name and Address, and Employee has Hire Date and Department, and obviously PersonID referring back to Person table. So, 2 rows in FCO table (person and employee), 3 rows in Src table, 8 rows in Attribs.
The view, we'll call it vw_Employee, selects PersonID, Name, Address, Hire Date, Department from the two tables. It is built by a SQL stored procedure we'll call OnMetadataChange.
This SP is fired (by trigger or batch process), and its purpose is to generate the CREATE VIEW statements. It will iterate through every First Class Object, collect which fields from which tables constitute the view, and will issue a CREATE statement based on that. So OnMetadataChange produces a DROP and CREATE for each view, it generates a dynamic SQL statement that is executed once per entry in FCO table. It is preferable to do this with Triggers but not necessary. Hopefully your FCO definitions won't change too often, and when they do, there will probably be a code release as well. You can run your OnMetadataChange SP at that time.
The end result is a 2-layer database. The views constitute the First Class Object layer, which is meaningful to the application. The application only uses views. The tables constitute the 'physical' layer, which the application shouldn't care about. The meta-tables are essentially your mapping between the FCO layer and the physical layer. It takes some time to set it up, but it's quite effective, and gives you many of the benefits of EAV, while at the same time giving you the concrete benefits of 3nf tables (indexability, etc).
If you'd like I can throw some sample SQL out there.

Part of the problem you are having is that you are trying to store schema-less data in a SQL database, which is not its strength. There are three approaches that would make your life far easier:
1) Have a column which stores the serialized custom fields, with whatever format is mst convenient. For example, this column could store xml. Upsides are that you can use SQL Server Compact and pulling back a record is trivial. Downsides are that you always have to pull/push the entire xml blob to do an update, and it is difficult to impossible to query on any custom fields.
2) Upgrade to SQL Server Express, and use XML columns. This is nearly the same as the first suggestion, except that any server ready version of SQL Server has native support for XML data. These columns can have indexes added and fields within the data can be used in queries.
3) Use a Schema-less Database, like MongoDB or CouchDB. These databases are all about storing schemaless data, so your custom fields will be no different than any other field. As such, you can index and query custom fields. Upsides are that custom data is incredibly easy to work with, downsides are that you would have to spend some time rethinking how you store data to fit within their model.
If you do not need to query based on custom fields, or if you can query custom fields within business logic, then the first option can work for you. In any other case, I would err towards something with more capabilities than compact. If cost is the deciding factor, both SQL Server Express and MongoDB are free.

Storing Data from Forms without creating 100's of tables: ASP.NET and SQL Server

Let me first describe the situation. We host many Alumni events over the course of each year and provide online registration forms for each event. There is a large chunk of data that is common for each event:
An Event with dates, times, managers, internal billing info, etc.
A Registration record with info about the payment and total amount charged per form submission
Bio/Demographic and alumni data about the 1 or more attendees (name, address, degree, etc.)
We store all of the above data within columns in tables as you would expect.
The trouble comes with the 'extra' fields we are asked to put on the forms. Maybe it is a dinner and there is a Veggie or Carnivore option, perhaps there is lodging and there are bed or smoking options, or perhaps there is an optional transportation option. There are tons of weird little "can you add this to the form?" types of requests we receive.
Currently, we JSONify any non-standard data and store it all in one column (per attendee) called 'extras'. We can read this data out in code but it is not well suited to querying. Our internal staff would like to generate a quick report on Veggie dinners needed for instance.
Other than creating a separate table for each form that holds the specific 'extra' data items, are there any other approaches that could make my life (and reporting) easier? Anyone working in a simialr environment?

This is actually one of the toughest problem to solve efficiently. The SQL Server Customer Advisory Team has dedicated a white-paper to the topic which I highly recommend you read: Best Practices for Semantic Data Modeling for Performance and Scalability.
You basically have 3 options:
semantic database (entity-attribute-value)
XML column
sparse columns
Each solution comes with ups and downs. Out of the top of my hat I'd say XML is probably the one that gives you the best balance of power and flexibility, but the optimal solution really depends on lots of factors like data set sizes, frequency at which new attributes are created, the actual process (human operators) that create-populate-use these attributes etc, and not at least your team skill set (some might fare better with an EAV solution, some might fare better with an XML solution). If the attributes are created/managed under a central authority and adding new attributes is a reasonable rare event, then the sparse columns may be a better answer.

Well you could also have the following db structure:
Have a table to store custom attributes
AttributeID
AttributeName
Have a mapping table between events and attributes with:
AttributeID
EventID
AttributeValue
This means you will be able to store custom information per event. And you will be able to reuse your attributes. You can include some metadata as
AttributeType
AllowBlankValue
to the attribute to handle it easily afterwards

Have you considered using XML instead of JSON? Difference: XML is supported (special data type) and has query integration ;)

quick and dirty, but actually nice for querying: simply add new columns. it's not like the empty entries in the previous table should cost a lot.
more databasy solution: you'll have something like an event ID in your table. You can link this to an n:m table connecting events to additional fields. And then store the additional field data in a table with additional_field_id, record_id (from the original table) and the actual value. Probably creates ugly queries, but seems politically correct in terms of database design.
I understand "NoSQL" (not only sql ;) databases like couchdb let you store arbitrary fields per record, but since you're already with SQL Server, I guess that's not an option.

This is the solution that we first proposed in ASP.NET Forums (that later became Community Server), and that the ASP.NET team built a similar version of in the ASP.NET 2.0 Membership when they released it:
Property Bags on your domain objects
For example:
Event.Profile() or in your case, Event.Extras().
Basically, a property bag is a serialized collection of data stored in a name/value pair in a column (or columns). The ASP.NET 2.0 Membership went the route of storing names in a semi-colon delimited list, and values in the same:
Table: aspnet_Profile
Column: PropertyNames (separated by semi-colons, and has start index and end index)
Column: PropertyValues (separated by semi-colons, and only stores the string value)
The downside to that approach is it is all strings, and manually has to be parsed (even though the membership system does it for you automatically).
Recently, my current method is I've built FormCollection and NameValueCollection C# extension methods that automatically serialize the collections to an XML result. And I store that XML in the table in it's own column associated with that entity. I also have a deserializer C# extension on XElement that deserializes that data back to the collection at runtime.
This gives you the power of actually querying those properties in XML, via SQL (though, that can be slow though - always flatten out your read-only data).
The final note is runtime querying: The general rule we follow is, if you are going to query a property of an entity in normal application logic, then you move that property to an actual column on the table - and create the appropriate indexes. If that data will never be queried directly (for example, Linq-to-Sql or EF), then leave it in the XML Property Bag.
Property Bags gives you the power of extending your domain models however you like, without having to modify the db schema.