At the top of the SQLite library stack, according to documentation, much of the public interface to the SQLite library is implemented by the wen.c, legacy.c, and vdbeapi.c source files. This is the point of communication for other programs and scripts.

Tokenizer breaks the SQL string passed from the interface into tokens and hands the tokens over to the parser, one by one. Tokenizer is hand-coded in C. The parser for SQLite is generated by the Lemon parser generator. It is faster than YACC and Bison and, at the same time, is thread safe and prevents memory leaks. The parser builds a parse tree from the tokens passed by the tokenizer and passes the tree to the code generator. The generator produces virtual machine code from the input and passes it to the virtual machine as executables. More information about the Lemon parser generator can be found at http://en.wikipedia.org/wiki/Lemon_Parser_Generator.

The virtual machine, also known as Virtual Database Engine (VDBE), is the heart of SQLite. It is responsible for fetching and changing values in the database. It executes the program generated by the code generator to manipulate database files. Each SQL statement is first converted into virtual machine language for VDBE. Each instruction of VDBE contains an opcode and up to three additional operands.

B-trees, along with Pager and the OS Interface, form the backend of the SQLite architecture. B-trees are used to organize the data. The pager on the other hand assists B-tree by caching, modifying, and rolling back data. B-tree, when required, requests particular pages from the cache; this request is processed by the pager in an efficient and reliable manner. The OS Interface, as the name suggests, provides an abstraction layer to port to different operating systems. It hides the unnecessary details of communicating with different operating systems from SQLite calls and handles them on behalf of SQLite.

These are the internals of SQLite and an application developer in Android need not worry about the internals of Android because the SQLite Android libraries have effectively used the concept of abstraction and all the complexities are hidden. One just needs to master the APIs provided, and that will cater to all the possible use cases of SQLite in an Android application.