Re: Huge performance gap between iOS SQLite and Android SQLite
We use sqlcipher (based on SQLite) here in IBM have not heard horrors stories. Like Jesse said you could be running Android 1.6 :-) if you don't provide something to replicate hard to tell also stackoverflow might provide wider audience for help Also hardware specs and if io type internal or SD card might also be helpful On Tue, Jul 21, 2015 at 7:15 PM Parashuram N panar...@microsoft.com wrote: It would also be interesting to see the SQLite plugin that you are using for this test. -Original Message- From: Jesse [mailto:purplecabb...@gmail.com] Sent: Tuesday, July 21, 2015 3:33 PM To: dev@cordova.apache.org Subject: Re: Huge performance gap between iOS SQLite and Android SQLite Please post a brief sample demonstrating the issue and someone will have a look. It is virtually impossible without even knowing what plugin you are using Gut reaction is the android code is doing something wrong, but it's hard to know with a concrete example. My team is hiring! @purplecabbage risingj.com On Tue, Jul 21, 2015 at 8:12 AM, Rob Sherman rob.sher...@medicaltracking.com wrote Hello Cordova Dev's. My company develops a mobile inventory/ERP client/scheduling/Logistics App for the medical industry. We support iOS and Android (and possibly Windows Tablets soon) We are seeing (as are many others) extreme differences in SQLite data load times between iOS and Android. We have applied every optimization and tried nearly anything and everything on Stack overflow. A data load (lets call it an initial data load) for a large client with a lot of data takes 1h 57s 342ms on Android -the same code and same plug-in on iOS takes less than 0h 6m 112ms to do the same work. This is a radical difference, to be sure. I understand the plug-in developer is likely a better contact on this, however; 1. In the enterprise market, which is growing fast, local DB can be considered a core competancy. 2. After applying every possible technique, including not transaction wrapping each insert-the gains are marginal. 3. This seems to indicate a platform difference better addressed centrally We are getting creative in that we are trying locking preferences and even waiting to apply triggers until after init data load -this is in development for POC and dev testing now but I haven't had result reports to review yet, so I can't detail the effectiveness, yet. Any help/direction and/or action is sincerely appreciated-and by no means expected, I am guessing you are at least aware there is difference. Hopefully my voice added to other might spur an organization wide discussion that leads to equality of SQLite performance across platforms as the very nature of Cordova and the HTML5/JavaScript components manage to do achieve parity. Sincerely, Rob Sherman, Mobile Architect, MTS
RE: Huge performance gap between iOS SQLite and Android SQLite
It would also be interesting to see the SQLite plugin that you are using for this test. -Original Message- From: Jesse [mailto:purplecabb...@gmail.com] Sent: Tuesday, July 21, 2015 3:33 PM To: dev@cordova.apache.org Subject: Re: Huge performance gap between iOS SQLite and Android SQLite Please post a brief sample demonstrating the issue and someone will have a look. It is virtually impossible without even knowing what plugin you are using Gut reaction is the android code is doing something wrong, but it's hard to know with a concrete example. My team is hiring! @purplecabbage risingj.com On Tue, Jul 21, 2015 at 8:12 AM, Rob Sherman rob.sher...@medicaltracking.com wrote Hello Cordova Dev's. My company develops a mobile inventory/ERP client/scheduling/Logistics App for the medical industry. We support iOS and Android (and possibly Windows Tablets soon) We are seeing (as are many others) extreme differences in SQLite data load times between iOS and Android. We have applied every optimization and tried nearly anything and everything on Stack overflow. A data load (lets call it an initial data load) for a large client with a lot of data takes 1h 57s 342ms on Android -the same code and same plug-in on iOS takes less than 0h 6m 112ms to do the same work. This is a radical difference, to be sure. I understand the plug-in developer is likely a better contact on this, however; 1. In the enterprise market, which is growing fast, local DB can be considered a core competancy. 2. After applying every possible technique, including not transaction wrapping each insert-the gains are marginal. 3. This seems to indicate a platform difference better addressed centrally We are getting creative in that we are trying locking preferences and even waiting to apply triggers until after init data load -this is in development for POC and dev testing now but I haven't had result reports to review yet, so I can't detail the effectiveness, yet. Any help/direction and/or action is sincerely appreciated-and by no means expected, I am guessing you are at least aware there is difference. Hopefully my voice added to other might spur an organization wide discussion that leads to equality of SQLite performance across platforms as the very nature of Cordova and the HTML5/JavaScript components manage to do achieve parity. Sincerely, Rob Sherman, Mobile Architect, MTS
Re: Huge performance gap between iOS SQLite and Android SQLite
Please post a brief sample demonstrating the issue and someone will have a look. It is virtually impossible without even knowing what plugin you are using Gut reaction is the android code is doing something wrong, but it's hard to know with a concrete example. My team is hiring! @purplecabbage risingj.com On Tue, Jul 21, 2015 at 8:12 AM, Rob Sherman rob.sher...@medicaltracking.com wrote: Hello Cordova Dev's. My company develops a mobile inventory/ERP client/scheduling/Logistics App for the medical industry. We support iOS and Android (and possibly Windows Tablets soon) We are seeing (as are many others) extreme differences in SQLite data load times between iOS and Android. We have applied every optimization and tried nearly anything and everything on Stack overflow. A data load (lets call it an initial data load) for a large client with a lot of data takes 1h 57s 342ms on Android -the same code and same plug-in on iOS takes less than 0h 6m 112ms to do the same work. This is a radical difference, to be sure. I understand the plug-in developer is likely a better contact on this, however; 1. In the enterprise market, which is growing fast, local DB can be considered a core competancy. 2. After applying every possible technique, including not transaction wrapping each insert-the gains are marginal. 3. This seems to indicate a platform difference better addressed centrally We are getting creative in that we are trying locking preferences and even waiting to apply triggers until after init data load -this is in development for POC and dev testing now but I haven't had result reports to review yet, so I can't detail the effectiveness, yet. Any help/direction and/or action is sincerely appreciated-and by no means expected, I am guessing you are at least aware there is difference. Hopefully my voice added to other might spur an organization wide discussion that leads to equality of SQLite performance across platforms as the very nature of Cordova and the HTML5/JavaScript components manage to do achieve parity. Sincerely, Rob Sherman, Mobile Architect, MTS
