Dynamic SQL in MemSQL - singlestore

We are considering using MemSQL to provide screening capabilities against our data. A screening operation can have an arbitrary number of criteria -- the type of requirement that would most likely be addressed using "Dynamic SQL" in Oracle or MySQL.
Would this type of arbitrary screening be a problem for MemSQL? This FAQ from MemSQL suggests that compiling a query takes time. Does it mean that each time MemSQL sees a new WHERE clause, it will have to compile a new C++ fragment? If so, how long of a delay would that compilation introduce?
Thanks.

That is correct, as described in the FAQ each new WHERE clause will require query compilation, unless the only difference is the values of parameters. (This is because we compile and optimize the plan based on the filters available in the where clause.) So if you add a new filter, that would require another compilation. In the current release of memsql, compilation for a select query may take ~1s or more, depending on the query.
One workaround to avoid this, if you know all possible criteria you're interested in beforehand, is to write a WHERE clause like:
WHERE (a=1 OR 0) AND (b<5 OR 0) AND (c=3 OR 0)
which is equivalent to
WHERE (a=1) AND (b<5) AND (c=3)
Of course, those filters can be arbitrary expressions. By changing an OR 0 to OR 1, you ignore that condition, so e.g. to get just b<5 you write
WHERE (a=1 OR 1) AND (b<5 OR 0) AND (c=3 OR 1)
This doesn't require additional compilation, because only the parameters are changing. Note however that it may not perform as well because the compiled query plan has to check each of those clauses at runtime (that's how this workaround avoids recompilation) and the best plan may be different depending on which filters are actually present.

Related

Customize jOOQ dialect to alter the order in which LIMIT and OFFSET are rendered in a statement

I'm using jOOQ to generate queries to run against Athena (AKA PrestoDB/Trino)
To do this, I am using SQLDialects.DEFAULT, and it works because I use very basic query functionalities.
However, jOOQ renders queries like this:
select *
from "Artist"
limit 10
offset 10
God knows why, but the order of limit and offset seem to matter, and the query only works if written with the order swapped:
select *
from "Artist"
offset 10
limit 10
Is there a class I can subclass, to modify the statement render function so that the order of these are swapped? Or any other way of implementing this myself?
A generic solution in jOOQ
There isn't a simple way to change something as fundamental as the SELECT clause order (or any other SELECT clause syntax) so easily in jOOQ, simply, because this was never a requirement for core jOOQ usage, other than supporting fringe SQL dialects. Since the support of a SQL dialect is a lot of work in jOOQ (with all the integration tests, edge cases, etc.) and since market shares of those dialects are low, it has simply never been a priority to improve this in jOOQ.
You may be tempted to think that this is "just" about the order of keywords in this one case. "Only this one case." It never is. It never stops, and the subtle differences in dialects never end. Just look at the jOOQ code base to get an idea of how weirdly different vendors choose to make their dialects. In this particular case, one would think that it seems extremely obvious and simple to make this clause MySQL / PostgreSQL / SQLite compatible, so your best chance is to make a case with the vendor for a feature request. It should be in their own best interest to be more compatible with the market leaders, to facilitate migration.
Workarounds in jOOQ
You can, of course, patch your generated SQL on a low level, e.g. using an ExecuteListener and a simple regex. Whenever you encounter limit (\d+|\?) offset (\d+|\?), just swap the values (and bind values!). This might work reasonably well for top level selects. It's obviously harder if you're using LIMIT .. OFFSET in nested selects, but probably still doable.
Patching jOOQ is always an option. The class you're looking for in jOOQ 3.17 is org.jooq.impl.Limit. It contains all the rendering logic for this clause. If that's your only patch, then it might be possible to upgrade jOOQ. But obviously, patching is a slippery slope, as you may start patching all sorts of clauses, making upgrades impossible.
You can obviously use plain SQL templates for simple cases, e.g. resultQuery("{0} offset {1} limit {2}", actualSelect, val(10), val(10)). This doesn't scale well, but if it's only about 1-2 queries, it might suffice
Using the SQLDialect.DEFAULT
I must warn you, at this point, that the behaviour of SQLDialect.DEFAULT is unspecified. Its main purpose is to produce something when you call QueryPart.toString() on a QueryPart that is not an Attachable, where a better SQLDialect is unavailable. The DEFAULT dialect may change between minor releases (or even patch releases, if there's an important bug in some toString() method), so any implementation you base on this is at risk of breaking with every upgrade.
The most viable long term solution
... would be to have support for these dialects in jOOQ:
#5414 Presto
#11485 Trino

Pagination in Gremlin

I'm trying to do pagination in Gremlin. I've followed the solution provided on gremlin recipes. So right now I am using the range() step to truncate results. This works well because when users asks for x results id only queries for x of them and doesn't perform full search which is significantly faster.
However in Gremlin docs it states that:
A Traversal’s result are never ordered unless explicitly by means of order()-step. Thus, never rely on the iteration order between TinkerPop3 releases and even within a release (as traversal optimizations may alter the flow).
And order is really important for pagination (we don't want user to have same results on different pages).
Adding order() would really slow down querying, since it will have to query all vertices that apply to the search and then truncate it with range().
Any ideas how this can be solved to have consistency between queries and still have smaller query time for not-full search?
The documentation is correct, but could use some additional context. Without order() you aren't guaranteed a specific order, unless the underlying graph database guarantees that order. TinkerPop should preserve that guarantee if it exists. So, ultimately you need to consider what the underlying graph will do.
As you are using JanusGraph, I'm pretty sure that it does not guarantee order of result iteration but there are caveats depending on the type of indexing that you do. You can read more about that here but may want to ask specific questions on the JanusGraph user list.

"ignoreDocumentNotFound", "readCompleteInput" query options in explain

When I use "explain" on an insert query I get two query options that seems not to be documented:
ignoreDocumentNotFound
readCompleteInput
What are these options for and what they do?
Nice to see you like our db._explain() Facility ;-)
To answer your question one has to know that explain re-uses a backend functionality that is also used for different purposes:
distribute AQL queries in ArangoDB clusters
analyse what the optimizer did with queries in Unittests
The later will explain queries, and check whether certain assumptions over the query plan are still valid.
The ignoreDocumentNotFound and readCompleteInput flags are exactly intended for that purpose, so the unittests can revalidate whether certain assumptions for the query are still true.
Since they don't contain additional value for the end user, they're not documented. One could argue whether explain should hide them to avoid irretations

What code could be used as a string aggregator for Sybase? (Like Oracle's stragg)

In my travels in Oracle, the 'stragg' function, or 'String Aggregator' was life-saving when I had to create dynamic SQL queries on the fly.
You can read up about it here: http://www.oratechinfo.co.uk/delimited_lists_to_collections.html
The basic use of it was:
select stragg(fruit) from food;
fruit
-----------
apple,pear,banana,strawberry
1 row(s) returned
So simple to use, concatenating chr(13) turned it into a long list, and selecting information from system tables gave a 5 minute solution to dynamically generated SQL, e.g. auditing triggers.
Now I've been charged with transferring oracle functionality related to auditing into Sybase, and a function similar to Stragg would be ideal for this purpose.
E.g.
select #my_table = 'table_of_fruit'
select 'insert into '+#mytable+'_copy (' +char(10)
+ stragg(c.name) +char(10)
+ 'select '
+ stragg('inserted.'+c.name) + char(10)
+ 'from '+#mytable
from syscolumns c
where objectid(#mytable) = c.id
------------------------------------------
insert into table_of_fruit_copy
(fruit, sweetness, price)
select fruit, sweetness,price
from inserted
Done. Simple.
Except I don't know how to get a string-aggregation function working in Sybase.
Does anyone know of an attempt to do this kind of thing, or code that could work the same as stragg that could be used in this way?
The alternative at the moment is printing code based on complex cursors and such (sample LOC: 500), or select statements combining static strings and columns from user tables (sample LOC: 200). Stragg would severely reduce the complexity of this code, and would be a great deal of help in the future (sample LOC: who knows, maybe 50?)
p.s. I'm calling these selects through a shell script then piping them to file, then running the file through iSQL. Not the nicest solution, but it's better than the alternatives.
There are three separate answers
Question
You have made comments about simplicity, which need to be addressed before we get to the solution.
It is a common requirement to be able to take a delimited list of values, say A,B,C,D, and treat this data like it was a set of rows in a table, or vice versa
This one of the Top Ten Worst Programming Practices I read about recently.
In general, Sybase types tend to be somewhat more academically and Relationally qualified than Oracle types, so we simply do not do that sort of thing in SybaseLand or DB2Land.
In 20 years of working with Sybase, I have had to code that as part of my project just once, and that was for non-technical Auditor who loaded the result set into MS Access.
On the other hand, I have had to code that at least 12 times, when producing text files for importation into Oracle databases (fulfilling external requirements is outside my project, but I satisfy any such requirement free). Obviously the target databases were sub-standard and non-relational (loading a column with more than one datum breaks 1NF, and creates Update Anomalies), which is typical of what Oracle types have to do to get some speed.
Therefore, no, it is not simplicity, at least in the sense of that principle. It is by definition, complexity.
Your reference to "arrays" is incorrect. All commercial dbms handle arrays, according to the ISO/IEC/ANSI SQL (STRAGGR and LIST operators are non-standard SQL, therefore not SQL). Sybase is very strong in processing arrays. If it was an array, you would not need special hand coding to handle it (and you do, as per your question). This is not an array, there is no definition to the cells. This is a single concatenated scalar string.
Pivoting is an entirely different process, which uses set-processing; it does not require row-processing. (I understand on good authority, that Oracle is hopeless at scalar subqueries, and thus Oracle people are used to writing them as [very inefficient] joins or inline views, and then filtering: all that can be elevated to set-processing via scalar subqueries, and it will perform much faster. Particularly your Pivots.)
Even the author in your link posts as follows. Please familiarise yourself with the caveats:
It's as simple as this: If you want to have a system with no logical limitation in the number of data elements passed to a given process, then forget the following mechanisms! They are simply the wrong way to approach the problem.
Therefore, know whatever you are doing is sub-standard, non-relational, and limited; and go ahead with your eyes open. No use pretending that: it will not break; it is not limited; it is an "array"; or that Sybase doesn't have a neat little function that Oracle has. Any professional will see through all that. And if the string length is exceeded, for God's sake send some indicator back to the caller ("!Exceeded" in the string) identifying that condition.
Essentially you are turning the set-processing engine on its head, and forcing it into row-processing mode, so it will be very slow. A WHILE loop is distinctly faster than a cursor, but both are in the same class, row-processors.
The alternative at the moment is printing code based on complex cursors and such
What 200 or 500 LoC ? It is possible I am missing something, but my code is the same few lines of code identified under "Using a Table Function" in your link. Maximum 20, if you count nice formatting; the loop; initialisation; error handling. There is nothing "complex" about it. Do the exact reverse to cancatenate a single string from multiple rows. We use stored procedures for this (which oracle does not have, really, PL/SQL is a different animal). If you have ASE 15.0.2 or greater, you can use a User Defined Function, which you can then use in place of a column. Stored procs are better for true arrays.
the concatenation operator in Sybase is the plus sign. For reversal (decomposing the CSV string) you need CHARINDEX and SUBSTRING functions
You may need the Function Reference Manual, if for nothing else, to avoid writing code where we have functions.
Likewise, we do not have a RANK() function. We are quite happy with the 4 lines of code requires for the subquery. It is only required for Oracle because subqueries are crippled.
Ok, I have answered your question, Now to address the approaches.
You will be aware that code using Oracle Extensions to the SQL standard will need to be changed.
Sybase is way more automated than Oracle; if you familiarise yourself with its feature set, in many instances, you can get the same result (as you did in Oracle) without writing any code. Writing code-for-code blocks is the chain gang, rock-breaking method of building roads, in the context of bulldozers. Even if your company had good reason to use that method, you need to the aware that features work quite differently, eg. triggers, which is why I am posting so much detail.
Another issue that will annoy you is that Oracle isn't really ANSI SQL compliant (stretches the definitions in many places, in order to appear to be compliant), and Sybase, given its customer base, is rigidly SQL compliant. So in addition to the same function working differently, or in a different deployment, you need to be aware that code changes may be required to elevate Oracle code to ANSI compliance levels, just to execute on an ANSI SQL compliant platform.
I am not sure if you are trying to write code for the content of a trigger, or if you are trying to capture the changes to a database. I will provide both answers.
Auditing
Capture Changes to Database
We have an very robust, fast and configurable Auditing subsystem, fit for high volumes and banking level auditing requirements. Get your DBA to setup the sybaudit (separate) database, and to configure exactly what changes need to be captured. This facility will perform much faster than any code you or I can write in a trigger (as much as 100 times faster than your row-by-processing required for the above, as it is executed within the engine, within your executing thread). And of course the setup time is a fraction of your coding time.
Triggers
Again, I am not sure exactly what you are trying to achieve, but assuming you want to copy every insert to some table to a COPY of that table (inside the Trigger), that example code you have provided will not work (and I am not counting syntax issues).
Speaking to your example, you need to do way more work, to deal with the different datatypes; column sizes; precisions; scale; etc. And perhaps the UPDATE() function to identify which columns have changed (for an UPDATE trigger of course). If all you are trying to do is convert the various datatypes to strings, check the CONVERT() function.
Triggers are transactional.
Never place row-processing code in a Trigger (it will strangle the table)
You can't place Dynamic SQL in a Trigger.
But in Sybase even that is not necessary. Refer to the User Guide, chapter 19 is devoted to Triggers, with several variations, and examples. Inside the trigger, you should be able to simply:
INSERT table_copy
SELECT column_list -- never use * unless you want the db fixed in cement
FROM inserted
If you are trying to copy the inserts to all tables into one Audit table, then beware. Then I understand your example a little bit more. You will be forcing a highly Symmetric Muli-Threading server (oracle is not a server in the architecture sense) into single-threading through your table. Auditing is multi-threaded.
Last, the use of manual methods of any kind is not required, so if you could expand a bit more on your PS, what the requirement you are trying to fulfil is, I can identify the programmatic method for you. It appears you are trying to use the PL/SQL approach (which is very limited).
Just use the LIST() function. It's a direct replacement for stragg() function. Example:
SELECT LIST(state, ', ') FROM cities
Result:
name
CA, CA, MA, NY

How do I sort Lucene results by field value using a HitCollector?

I'm using the following code to execute a query in Lucene.Net
var collector = new GroupingHitCollector(searcher.GetIndexReader());
searcher.Search(myQuery, collector);
resultsCount = collector.Hits.Count;
How do I sort these search results based on a field?
Update
Thanks for your answer. I had tried using TopFieldDocCollector but I got an error saying, "value is too small or too large" when i passed 5000 as numHits argument value. Please suggest a valid value to pass.
The search.Searcher.search method will accept a search.Sort parameter, which can be constructed as simply as:
new Sort("my_sort_field")
However, there are some limitations on which fields can be sorted on - they need to be indexed but not tokenized, and the values convertible to Strings, Floats or Integers.
Lucene in Action covers all of the details, as well as sorting by multiple fields and so on.
What you're looking for is probably TopFieldDocCollector. Use it instead of the GroupingHitCollector (what is that?), or inside it.
Comment on this if you need more info. I'll be happy to help.
In the original (Java) version of Lucene, there is no hard restriction on the size of the the TopFieldDocCollector results. Any number greater than zero is accepted. Although memory constraints and performance degradation create a practical limit that depends on your environment, 5000 hits is trivial and shouldn't pose a problem outside of a mobile device.
Perhaps in porting Lucene, TopFieldDocCollector was modified to use something other than Lucene's "heap" implementation (called PriorityQueue, extended by FieldSortedHitQueue)—something that imposes an unreasonably small limit on the results size. If so, you might want to look at the source code for TopFieldDocCollector, and implement your own similar hit collector using a better heap implementation.
I have to ask, however, why are you trying to collect 5000 results? No user in an interactive application is going to want to see that many. I figure that users willing to look at 200 results are rare, but double it to 400 just as factor of safety. Depending on the application, limiting the result size can hamper malicious screen scrapers and mitigate denial-of-service attacks too.
The constructor for Sort accepting only the string field name has been depreciated. Now you have to create a sort object and pass it in as the last paramater of searcher.Search()
/* sorting by a field of type long called "size" from greatest -> smallest
(signified by passing in true for the last isReversed paramater)*/
Sort sorter = new Sorter(new SortField("size", SortField.Type.LONG, true))
searcher.Search(myQuery, collector, sorter);

Resources