[android-developers] Re: How does a Cursor work?
Thanks Satya, you did help me. After reading your email I looked at CursorWindow, and I could see more clearly how SQLiteCursor works, including the native part of CursorWindow implementation. I could see the notion of windowing is very interesting and fast, mainly its native part. As I could see, a window keeps only a necessary number of rows. I haven't gone deep enough to see what is considered a necessary number of rows, but I *believe* the implementation employs some kind of spatial locality, usually employed when accessing memory address... Thank you, Satya, Taísa On Thu, Dec 11, 2008 at 2:45 PM, Satya Komatineni satya.komatin...@gmail.com wrote: Taisa, Hopefully you have found some answers on this since you have posted. If you did find any numbers indicating one way or the other, I would like to know. I looked at some source code of Android to see what is under the hood. Here are some thoughts based on what I saw. The Cursor object is an interface that is allowing both forward and backward movement. In addition the interface also supports the getCount(). Both seem to indicate that all rows might be read upfront or at the earliest moment. The implementation of this cursor interface using SQLiteCursor seem to be using a windowing concept to read a set of rows depending on the window you are in. So it is possible that moving forward in a cursor should be pretty efficient. However calling getCount() may force a complete read of the internal cursor. So semantics of the Cursor interface is not delineating a clear forward only and random access semantics. But I believe the implementation is efficient enough if you follow the forward only semantics. This implies reading getCount early on is not a good thing if you are trying to read every row and do something with it. It may not be a bad idea to wrap the cursor interface and don't expose the getCount() and random movement methods and make the contract explicit to the clients. Hope that helps. On Mon, Dec 8, 2008 at 1:51 PM, Taísa Cristina taisa.san...@gmail.com wrote: Hi, when a database query retrieves a Cursor, what does it have in fact? I mean, does it have the whole result set in memory or keep a kind of pointer to each result row, and when I do cursor.moveToNext() it points to the next row? Or anything else? I need to deal with a long list of data, and I really wanna know how efficient is a Cursor retrieved from a database query. Thanks, Taísa -- Taísa Cristina Costa dos Santos Computer Engineer Brazil, SP 55 19 8152-7453 --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Android Developers group. To post to this group, send email to android-developers@googlegroups.com To unsubscribe from this group, send email to android-developers-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/android-developers?hl=en -~--~~~~--~~--~--~---
AW: [android-developers] Re: How does a Cursor work?
hi satya, great comments! thanx for the information. it might be a clever alternative to implement the count(*) functionality by means of a prepared (compiled) statement, if you need the number of results. the cursor will do the iteration job, the prep statement the fast initial counting. marcus Von: Satya Komatineni satya.komatin...@gmail.com An: android-develop...@googlegroups..com Gesendet: Donnerstag, den 11. Dezember 2008, 17:45:32 Uhr Betreff: [android-developers] Re: How does a Cursor work? Taisa, Hopefully you have found some answers on this since you have posted. If you did find any numbers indicating one way or the other, I would like to know. I looked at some source code of Android to see what is under the hood. Here are some thoughts based on what I saw. The Cursor object is an interface that is allowing both forward and backward movement. In addition the interface also supports the getCount(). Both seem to indicate that all rows might be read upfront or at the earliest moment. The implementation of this cursor interface using SQLiteCursor seem to be using a windowing concept to read a set of rows depending on the window you are in. So it is possible that moving forward in a cursor should be pretty efficient. However calling getCount() may force a complete read of the internal cursor. So semantics of the Cursor interface is not delineating a clear forward only and random access semantics. But I believe the implementation is efficient enough if you follow the forward only semantics. This implies reading getCount early on is not a good thing if you are trying to read every row and do something with it. It may not be a bad idea to wrap the cursor interface and don't expose the getCount() and random movement methods and make the contract explicit to the clients. Hope that helps. On Mon, Dec 8, 2008 at 1:51 PM, Taísa Cristina taisa.san...@gmail.com wrote: Hi, when a database query retrieves a Cursor, what does it have in fact? I mean, does it have the whole result set in memory or keep a kind of pointer to each result row, and when I do cursor.moveToNext() it points to the next row? Or anything else? I need to deal with a long list of data, and I really wanna know how efficient is a Cursor retrieved from a database query. Thanks, Taísa --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Android Developers group. To post to this group, send email to android-developers@googlegroups.com To unsubscribe from this group, send email to android-developers-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/android-developers?hl=en -~--~~~~--~~--~--~---
[android-developers] Re: How does a Cursor work?
Taisa, Hopefully you have found some answers on this since you have posted. If you did find any numbers indicating one way or the other, I would like to know. I looked at some source code of Android to see what is under the hood. Here are some thoughts based on what I saw. The Cursor object is an interface that is allowing both forward and backward movement. In addition the interface also supports the getCount(). Both seem to indicate that all rows might be read upfront or at the earliest moment. The implementation of this cursor interface using SQLiteCursor seem to be using a windowing concept to read a set of rows depending on the window you are in. So it is possible that moving forward in a cursor should be pretty efficient. However calling getCount() may force a complete read of the internal cursor. So semantics of the Cursor interface is not delineating a clear forward only and random access semantics. But I believe the implementation is efficient enough if you follow the forward only semantics. This implies reading getCount early on is not a good thing if you are trying to read every row and do something with it. It may not be a bad idea to wrap the cursor interface and don't expose the getCount() and random movement methods and make the contract explicit to the clients. Hope that helps. On Mon, Dec 8, 2008 at 1:51 PM, Taísa Cristina [EMAIL PROTECTED] wrote: Hi, when a database query retrieves a Cursor, what does it have in fact? I mean, does it have the whole result set in memory or keep a kind of pointer to each result row, and when I do cursor.moveToNext() it points to the next row? Or anything else? I need to deal with a long list of data, and I really wanna know how efficient is a Cursor retrieved from a database query. Thanks, Taísa --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Android Developers group. To post to this group, send email to android-developers@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/android-developers?hl=en -~--~~~~--~~--~--~---
