● Scalar and Vector files ● Contents: ○ Run description ○ Scalar file: (module, scalar, value), histograms, result attributes ○ Vector file: vector data: (module, vectorname, vector data = (timestamp+value)*) Comparison: SQLite vs Text ● File size: about the same or a bit smaller than the text based format ● Vector Recording performance: 2-2.5x slower (net writing speed) ● Read performance: Depends on the complexity to get the data inside R. (R already has SQLite access library) Possible drawbacks: ● Speed ○ vector recording performance is about 2-2.5x slower than text-based file format ○ in actual simulations

"row_names". If FALSE, never translate row names. If NA, translate rownames only if they’re a character vector. A string is equivalent to TRUE, but allows you to override the default name. For backward compatibility to SQLite. ... Needed for compatibility with generic. Otherwise ignored. field.types character vector of named SQL field types where the names are the names of new table’s columns. If missing, types if and only if the first row has one fewer field that the number of columns. colClasses Character vector of R type names, used to override defaults when imputing classes from on-disk file. 8 initExtension