'Not using SQL directly' means you create a series data access interface like an ISeries interface that browse in a cursor-style through your series data, then implement it as a concrete class like DatabaseSeries that does the SQL job for you.
Talking about scaling issues means that you could do some kind of data manipulation best on processing by itself instead of executing SQL statements (example: an data analysis study that take a series and output 3 new series with heterogeneous values, based on first data series, you can calculate the resulting 3 series fetching data one time when you iterate through you ISeries interface; executing 3 SQL statements will perform 3 read cycles, you can perform fast and also stop wasting CPU resources). []'s -----Original Message----- From: sqlite-users-boun...@sqlite.org [mailto:sqlite-users-boun...@sqlite.org] On Behalf Of Rick Ratchford Sent: sexta-feira, 10 de julho de 2009 23:10 To: 'General Discussion of SQLite Database' Subject: Re: [sqlite] Is it Possible in SQL... #>I was trying to figuring out if you are doing something of #>graph data analysis, I do it almost everyday in our Stock #>Trader applications... #>I never did this way (direct SQL), cause our graph series #>data sources are implement throught a common interface, that #>could be a SQL query, a stream, a XML, whatever. #> #>Just a tip: implementing specific and well designed #>interfaces for series data manipulation should be the right #>way for you, avoid scaling issues cause of possible SQL #>limitations in the ways those series can/should be #>manipulated in some cases. This specific application requires very little user input. The user selects a market (stock/futures) from a list of available data files and that is it. The program then performs all kinds of different things on the dataset selected and for the most part presents its results in the form of values. There is one procedure, however, that will produce a 'graph' if the user clicks on a button. That's pretty much it. When you said you never use 'direct SQL', are you saying that you never use SQL that is hard coded in your program? If so, perhaps in the case of my application requiring virtually no interaction that it is acceptable for some of the internal procedures? What are "scaling issues"? Thanks. :-) Rick _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users