Anyone an idea?
The issue is: I am writing a high performance application. It has a SQL database which I use for persistence. In memory objects get updated, then the changes queued for a disc write (which is pretty much always an insert in a versioned table). The small time risk is given as accepted - in case of a crash, program code will resynclocal state with external systems.
Now, quite often I need to run lookups on certain values, and it would be nice to have standard interface. Basically a bag of objects, but with the ability to run queries efficiently against an in memory index. For example I have a table of "instruments" which all have a unique code, and I need to look up this code.... about 30.000 times per second as I get updates for every instrument.
Anyone an idea for a decent high performance library for this?
You should be able to use an in-memory SQLite database (:memory) with System.Data.SQLite.
Related
I'm developing an RPG in Dart, and I'm going to use IndexedDB for data persistence.
I will have two databases: one for read-only access and one for read-write access where save games will be stored. I was just wondering if I should read required data directly from the database or cache it in Maps. I could potentially have several hundred records that need to be pulled from the read-only database (enemies, game maps etc.) and I though pulling everything from the database may be less efficient than using Dart's Maps.
Oh, also each database will be stored in a map. Object Stores will be nested maps inside that map.
Should I read directly from the database, or should I put everything into a Map and read from there?
EDIT: Forgot to mention, the read-only database will be initialised with data from a JSON file located on the user's machine, not through AJAX.
I am confident that hundreds of records will present you no issue in IndexedDB. IDB was designed with that kind of scale in mind, and its async APIs -- while vexing for novices -- make sure your app stays responsive by design.
I am working on a demo designed to push IDB further than it should go, and have some easy-to-reach statistics for you. These are gets on a single index in a single store on a database.
Gets are blazing fast in IndexedDB. The issue with IDB at scale is typically writes.
One thousand success callbacks, one complete callback, were sub-second:
Ten thousand success callbacks, one complete callback, was about 5 seconds:
More than fifty thousand success callbacks fired in less than a minute:
Writes are much slower - bursty at first, but then slow after minutes and dog slow after hours. That's with any schema, but you'd likely have multiple indexes on location (both latitude and longitude at least, I imagine) so your writes will be especially slow (more indexes, more work to do to main those in inserts and updates).
Layout for the stats above (just as important as the stats themselves, make sure to design your schema according to how you need to access it):
I would go with direct database access and then monitor the performance and then optimize where noteable gains are to be expected. Premature optimization is seldom a good idea.
I’ve been tasked with optimizing a rather nasty stored procedure in a legacy system. It’s a database dedicated to search, and a new copy is being generate every day, with a lot of complex joins being de-normalized. No writes are being performed, only SELECTs, so I figured some easy improvements could be made by making the whole database read-only and changing the recovery model to “Simple”.
Much to my surprise, this didn’t help – at all! The stored procedure still takes the same amount of time of complete. If fact, I’m so surprised that I figured I did it wrong!
My questions:
Do I need to do anything other than setting “Database read-only” to “true”?
Am I wrong to expect significant performance improvement by making the database read-only?
Same for the recovery model: Shouldn’t “Simple” have some noticeable impact?
Are there other similar database-wide configurations that can improve performance in this scenario?
The stored procedure is huge, with temporary tables, 40+ tables joined in 20+ queries. But I’d like to optimize the database itself before I edit this proc.
Since no writes are performed by your SP, there is no reason to expect noticable performance improvement from changing recovery model and read-write mode.
As others mentioned, you should look into the query plan and optimize your queries.
Another hint: indexes in the database might get fragmented while the database is filled up. Since the data is not going to be modified any more, it might help to rebuild all the indexes with fillfactor 100 - this might help to get rid of fragmentation and to compact data.
Call this for each table in the database: ALTER INDEX ALL ON table_name REBUILD WITH (FILLFACTOR = 100).
Generally, I won't expect much of performance improvement from this, but it depends on the particular database.
Speaking of query optimization, there are very useful features in SQL Server 2005 and later: Execution Related and Index-Related Dynamic Management Views. In particular, sys.dm_exec_query_stats and missing indexes are of interest.
These give you almost the same information as Tuning Advisor, but using you real-life workload, so you don't need to simulate it and feed to the Advisor.
Have you tried using the Database Engine Tuning Advisor included in SQL Server? It will analyze your query and suggest new indexes that will improve the performance of the query. Some of them will be good, some will be bad (for example, I've seen it suggest adding every column in a table to an index, sometimes like 30 of them!), so I don't follow it blindly. Generally I'll add a few indexes and then retest, to find the suggestions that are the most important. I've used it to optimize many queries that I thought I had properly indexed, only to find I could get a lot more performance out of them.
I had a similar setup, large stored procedures with lots of large temp tables.
Our problem was that the joins with and between the temp tables was very slow.
I recommend that you look at your execution plan and try to add relevant indexes to the temp tables too if you have not already.
Looking for a bit of advice on how to optimise one of our projects. We have a ASP.NET/C# system that retrieves data from a SQL2008 data and presents it on a DevExpress ASPxGridView. The data that's retrieved can come from one of a number of databases - all of which are slightly different and are being added and removed regularly. The user is presented with a list of live "companies", and the data is retrieved from the corresponding database.
At the moment, data is being retrieved using a standard SqlDataSource and a dynamically-created SQL SELECT statement. There are a few JOINs in the statement, as well as optional WHERE constraints, again dynamically-created depending on the database and the user's permission level.
All of this works great (honest!), apart from performance. When it comes to some databases, there are several hundreds of thousands of rows, and retrieving and paging through the data is quite slow (the databases are already properly indexed). I've therefore been looking at ways of speeding the system up, and it seems to boil down to two choices: XPO or LINQ.
LINQ seems to be the popular choice, but I'm not sure how easy it will be to implement with a system that is so dynamic in nature - would I need to create "definitions" for each database that LINQ could access? I'm also a bit unsure about creating the LINQ queries dynamically too, although looking at a few examples that part at least seems doable.
XPO, on the other hand, seems to allow me to create a XPO Data Source on the fly. However, I can't find too much information on how to JOIN to other tables.
Can anyone offer any advice on which method - if any - is the best to try and retro-fit into this project? Or is the dynamic SQL model currently used fundamentally different from LINQ and XPO and best left alone?
Before you go and change the whole way that your app talks to the database, have you had a look at the following:
Run your code through a performance profiler (such as Redgate's performance profiler), the results are often surprising.
If you are constructing the SQL string on the fly, are you using .Net best practices such as String.Concat("str1", "str2") instead of "str1" + "str2". Remember, multiple small gains add up to big gains.
Have you thought about having a summary table or database that is periodically updated (say every 15 mins, you might need to run a service to update this data automatically.) so that you are only hitting one database. New connections to databases are quiet expensive.
Have you looked at the query plans for the SQL that you are running. Today, I moved a dynamically created SQL string to a sproc (only 1 param changed) and shaved 5-10 seconds off the running time (it was being called 100-10000 times depending on some conditions).
Just a warning if you do use LINQ. I have seen some developers who have decided to use LINQ write more inefficient code because they did not know what they are doing (pulling 36,000 records when they needed to check for 1 for example). This things are very easily overlooked.
Just something to get you started on and hopefully there is something there that you haven't thought of.
Cheers,
Stu
As far as I understand you are talking about so called server mode when all data manipulations are done on the DB server instead of them to the web server and processing them there. In this mode grid works very fast with data sources that can contain hundreds thousands records. If you want to use this mode, you should either create the corresponding LINQ classes or XPO classes. If you decide to use LINQ based server mode, the LINQServerModeDataSource provides the Selecting event which can be used to set a custom IQueryable and KeyExpression. I would suggest that you use LINQ in your application. I hope, this information will be helpful to you.
I guess there are two points where performance might be tweaked in this case. I'll assume that you're accessing the database directly rather than through some kind of secondary layer.
First, you don't say how you're displaying the data itself. If you're loading thousands of records into a grid, that will take time no matter how fast everything else is. Obviously the trick here is to show a subset of the data and allow the user to page, etc. If you're not doing this then that might be a good place to start.
Second, you say that the tables are properly indexed. If this is the case, and assuming that you're not loading 1,000 records into the page at once and retreiving only subsets at a time, then you should be OK.
But, if you're only doing an ExecuteQuery() against an SQL connection to get a dataset back I don't see how Linq or anything else will help you. I'd say that the problem is obviously on the DB side.
So to solve the problem with the database you need to profile the different SELECT statements you're running against it, examine the query plan and identify the places where things are slowing down. You might want to start by using the SQL Server Profiler, but if you have a good DBA, sometimes just looking at the query plan (which you can get from Management Studio) is usually enough.
I'm about to have to rewrite some rather old code using SQL Server's BULK INSERT command because the schema has changed, and it occurred to me that maybe I should think about switching to a stored procedure with a TVP instead, but I'm wondering what effect it might have on performance.
Some background information that might help explain why I'm asking this question:
The data actually comes in via a web service. The web service writes a text file to a shared folder on the database server which in turn performs a BULK INSERT. This process was originally implemented on SQL Server 2000, and at the time there was really no alternative other than chucking a few hundred INSERT statements at the server, which actually was the original process and was a performance disaster.
The data is bulk inserted into a permanent staging table and then merged into a much larger table (after which it is deleted from the staging table).
The amount of data to insert is "large", but not "huge" - usually a few hundred rows, maybe 5-10k rows tops in rare instances. Therefore my gut feeling is that BULK INSERT being a non-logged operation won't make that big a difference (but of course I'm not sure, hence the question).
The insertion is actually part of a much larger pipelined batch process and needs to happen many times in succession; therefore performance is critical.
The reasons I would like to replace the BULK INSERT with a TVP are:
Writing the text file over NetBIOS is probably already costing some time, and it's pretty gruesome from an architectural perspective.
I believe that the staging table can (and should) be eliminated. The main reason it's there is that the inserted data needs to be used for a couple of other updates at the same time of insertion, and it's far costlier to attempt the update from the massive production table than it is to use an almost-empty staging table. With a TVP, the parameter basically is the staging table, I can do anything I want with it before/after the main insert.
I could pretty much do away with dupe-checking, cleanup code, and all of the overhead associated with bulk inserts.
No need to worry about lock contention on the staging table or tempdb if the server gets a few of these transactions at once (we try to avoid it, but it happens).
I'm obviously going to profile this before putting anything into production, but I thought it might be a good idea to ask around first before I spend all that time, see if anybody has any stern warnings to issue about using TVPs for this purpose.
So - for anyone who's cozy enough with SQL Server 2008 to have tried or at least investigated this, what's the verdict? For inserts of, let's say, a few hundred to a few thousand rows, happening on a fairly frequent basis, do TVPs cut the mustard? Is there a significant difference in performance compared to bulk inserts?
Update: Now with 92% fewer question marks!
(AKA: Test Results)
The end result is now in production after what feels like a 36-stage deployment process. Both solutions were extensively tested:
Ripping out the shared-folder code and using the SqlBulkCopy class directly;
Switching to a Stored Procedure with TVPs.
Just so readers can get an idea of what exactly was tested, to allay any doubts as to the reliability of this data, here is a more detailed explanation of what this import process actually does:
Start with a temporal data sequence that is ordinarily about 20-50 data points (although it can sometimes be up a few hundred);
Do a whole bunch of crazy processing on it that's mostly independent of the database. This process is parallelized, so about 8-10 of the sequences in (1) are being processed at the same time. Each parallel process generates 3 additional sequences.
Take all 3 sequences and the original sequence and combine them into a batch.
Combine the batches from all 8-10 now-finished processing tasks into one big super-batch.
Import it using either the BULK INSERT strategy (see next step), or TVP strategy (skip to step 8).
Use the SqlBulkCopy class to dump the entire super-batch into 4 permanent staging tables.
Run a Stored Procedure that (a) performs a bunch of aggregation steps on 2 of the tables, including several JOIN conditions, and then (b) performs a MERGE on 6 production tables using both the aggregated and non-aggregated data. (Finished)
OR
Generate 4 DataTable objects containing the data to be merged; 3 of them contain CLR types which unfortunately aren't properly supported by ADO.NET TVPs, so they have to be shoved in as string representations, which hurts performance a bit.
Feed the TVPs to a Stored Procedure, which does essentially the same processing as (7), but directly with the received tables. (Finished)
The results were reasonably close, but the TVP approach ultimately performed better on average, even when the data exceeded 1000 rows by a small amount.
Note that this import process is run many thousands of times in succession, so it was very easy to get an average time simply by counting how many hours (yes, hours) it took to finish all of the merges.
Originally, an average merge took almost exactly 8 seconds to complete (under normal load). Removing the NetBIOS kludge and switching to SqlBulkCopy reduced the time to almost exactly 7 seconds. Switching to TVPs further reduced the time to 5.2 seconds per batch. That's a 35% improvement in throughput for a process whose running time is measured in hours - so not bad at all. It's also a ~25% improvement over SqlBulkCopy.
I am actually fairly confident that the true improvement was significantly more than this. During testing it became apparent that the final merge was no longer the critical path; instead, the Web Service that was doing all of the data processing was starting to buckle under the number of requests coming in. Neither the CPU nor the database I/O were really maxed out, and there was no significant locking activity. In some cases we were seeing a gap of a few idle seconds between successive merges. There was a slight gap, but much smaller (half a second or so) when using SqlBulkCopy. But I suppose that will become a tale for another day.
Conclusion: Table-Valued Parameters really do perform better than BULK INSERT operations for complex import+transform processes operating on mid-sized data sets.
I'd like to add one other point, just to assuage any apprehension on part of the folks who are pro-staging-tables. In a way, this entire service is one giant staging process. Every step of the process is heavily audited, so we don't need a staging table to determine why some particular merge failed (although in practice it almost never happens). All we have to do is set a debug flag in the service and it will break to the debugger or dump its data to a file instead of the database.
In other words, we already have more than enough insight into the process and don't need the safety of a staging table; the only reason we had the staging table in the first place was to avoid thrashing on all of the INSERT and UPDATE statements that we would have had to use otherwise. In the original process, the staging data only lived in the staging table for fractions of a second anyway, so it added no value in maintenance/maintainability terms.
Also note that we have not replaced every single BULK INSERT operation with TVPs. Several operations that deal with larger amounts of data and/or don't need to do anything special with the data other than throw it at the DB still use SqlBulkCopy. I am not suggesting that TVPs are a performance panacea, only that they succeeded over SqlBulkCopy in this specific instance involving several transforms between the initial staging and the final merge.
So there you have it. Point goes to TToni for finding the most relevant link, but I appreciate the other responses as well. Thanks again!
I don't really have experience with TVP yet, however there is an nice performance comparison chart vs. BULK INSERT in MSDN here.
They say that BULK INSERT has higher startup cost, but is faster thereafter. In a remote client scenario they draw the line at around 1000 rows (for "simple" server logic). Judging from their description I would say you should be fine with using TVP's. The performance hit - if any - is probably negligible and the architectural benefits seem very good.
Edit: On a side note you can avoid the server-local file and still use bulk copy by using the SqlBulkCopy object. Just populate a DataTable, and feed it into the "WriteToServer"-Method of an SqlBulkCopy instance. Easy to use, and very fast.
The chart mentioned with regards to the link provided in #TToni's answer needs to be taken in context. I am not sure how much actual research went into those recommendations (also note that the chart seems to only be available in the 2008 and 2008 R2 versions of that documentation).
On the other hand there is this whitepaper from the SQL Server Customer Advisory Team: Maximizing Throughput with TVP
I have been using TVPs since 2009 and have found, at least in my experience, that for anything other than simple insert into a destination table with no additional logic needs (which is rarely ever the case), then TVPs are typically the better option.
I tend to avoid staging tables as data validation should be done at the app layer. By using TVPs, that is easily accommodated and the TVP Table Variable in the stored procedure is, by its very nature, a localized staging table (hence no conflict with other processes running at the same time like you get when using a real table for staging).
Regarding the testing done in the Question, I think it could be shown to be even faster than what was originally found:
You should not be using a DataTable, unless your application has use for it outside of sending the values to the TVP. Using the IEnumerable<SqlDataRecord> interface is faster and uses less memory as you are not duplicating the collection in memory only to send it to the DB. I have this documented in the following places:
How can I insert 10 million records in the shortest time possible? (lots of extra info and links here as well)
Pass Dictionary<string,int> to Stored Procedure T-SQL
Streaming Data Into SQL Server 2008 From an Application (on SQLServerCentral.com ; free registration required)
TVPs are Table Variables and as such do not maintain statistics. Meaning, they report only having 1 row to the Query Optimizer. So, in your proc, either:
Use statement-level recompile on any queries using the TVP for anything other than a simple SELECT: OPTION (RECOMPILE)
Create a local temporary table (i.e. single #) and copy the contents of the TVP into the temp table
I think I'd still stick with a bulk insert approach. You may find that tempdb still gets hit using a TVP with a reasonable number of rows. This is my gut feeling, I can't say I've tested the performance of using TVP (I am interested in hearing others input too though)
You don't mention if you use .NET, but the approach that I've taken to optimise previous solutions was to do a bulk load of data using the SqlBulkCopy class - you don't need to write the data to a file first before loading, just give the SqlBulkCopy class (e.g.) a DataTable - that's the fastest way to insert data into the DB. 5-10K rows isn't much, I've used this for up to 750K rows. I suspect that in general, with a few hundred rows it wouldn't make a vast difference using a TVP. But scaling up would be limited IMHO.
Perhaps the new MERGE functionality in SQL 2008 would benefit you?
Also, if your existing staging table is a single table that is used for each instance of this process and you're worried about contention etc, have you considered creating a new "temporary" but physical staging table each time, then dropping it when it's finished with?
Note you can optimize the loading into this staging table, by populating it without any indexes. Then once populated, add any required indexes on at that point (FILLFACTOR=100 for optimal read performance, as at this point it will not be updated).
Staging tables are good! Really I wouldn't want to do it any other way. Why? Because data imports can change unexpectedly (And often in ways you can't foresee, like the time the columns were still called first name and last name but had the first name data in the last name column, for instance, to pick an example not at random.) Easy to research the problem with a staging table so you can see exactly what data was in the columns the import handled. Harder to find I think when you use an in memory table. I know a lot of people who do imports for a living as I do and all of them recommend using staging tables. I suspect there is a reason for this.
Further fixing a small schema change to a working process is easier and less time consuming than redesigning the process. If it is working and no one is willing to pay for hours to change it, then only fix what needs to be fixed due to the schema change. By changing the whole process, you introduce far more potential new bugs than by making a small change to an existing, tested working process.
And just how are you going to do away with all the data cleanup tasks? You may be doing them differently, but they still need to be done. Again, changing the process the way you describe is very risky.
Personally it sounds to me like you are just offended by using older techniques rather than getting the chance to play with new toys. You seem to have no real basis for wanting to change other than bulk insert is so 2000.
I have a daily batch process that involves selecting out a large number of records and formatting up a file to send to an external system. I also need to mark these records as sent so they are not transmitted again tomorrow.
In my naive JDBC way, I would prepare and execute a statement and then begin to loop through the recordset. As I only go forwards through the recordset there is no need for my application server to hold the whole result set in memory at one time. Groups of records can be feed across from the database server.
Now, lets say I'm using hibernate. Won't I endup with a bunch of objects representing the whole result set in memory at once?
Hibernate does also iterate over the result set so only one row is kept in memory. This is the default. If it to load greedily, you must tell it so.
Reasons to use Hibernate:
"Someone" was "creative" with the column names (PRXFC0315.XXFZZCC12)
The DB design is still in flux and/or you want one place where column names are mapped to Java.
You're using Hibernate anyway
You have complex queries and you're not fluent in SQL
Reasons not to use Hibernate:
The rest of your app is pure JDBC
You don't need any of the power of Hibernate
You have complex queries and you're fluent in SQL
You need a specific feature of your DB to make the SQL perform
Hibernate offers some possibilities to keep the session small.
You can use Query.scroll(), Criteria.scroll() for JDBC-like scrolling. You can use Session.evict(Object entity) to remove entities from the session. You can use a StatelessSession to suppress dirty-checking. And there are some more performance optimizations, see the Hibernate documentation.
Hibernate as any ORM framework is intended for developing and maintaining systems based on object oriented programming principal. But most of the databases are relational and not object oriented, so in any case ORM is always a trade off between convenient OOP programming and optimized/most effective DB access.
I wouldn't use ORM for specific isolated tasks, but rather as an overall architectural choice for application persistence layer.
In my opinion I would NOT use Hibernate, since it makes your application a whole lot bigger and less maintainable and you do not really have a chance of optimizing the generated sql-scripts in a quick way.
Furthermore you could use all the SQL functionality the JDBC-bridge supports and are not limited to the hibernate functionality. Another thing is that you have the limitations too that come along with each layer of legacy code.
But in the end it is a philosophical question and you should do it the way it fits you're way of thinking best.
If there are possible performance issues then stick with the JDBC code.
There are a number of well known pure SQL optimisations which
which would be very difficult to do in Hibernate.
Only select the columns you use! (No "select *" stuff ).
Keep the SQl as simple as possible. e.g. Dont include small reference tables like currency codes in the join. Instead load the currency table into memory and resolve currency descriptions with a program lookup.
Depending on the DBMS minor re-ordering of the SQL where predicates can have a major effect on performance.
If you are updateing/inserting only commit every 100 to 1000 updates. i.e. Do not commit every unit of work but keep some counter so you commit less often.
Take advantage of the aggregate functions of your database. If you want totals by DEPT code then do it in the SQL with " SUM(amount) ... GROUP BY DEPT ".