⨳ Shifty Metadata ⨳

2017 Jul 18, Tuesday

I’ve got a set of Excel files (.xlsx) containing weekly spend breakouts. These files are supposed to be formatted identically, but they’re not. Across these files, a couple numeric variables are sometimes read as text. That makes stacking data impossible.

I could probably develop some overly-complicated bit of code to extract the schema from the dictionary tables and then adjust the import code on the fly. But that’s work.

Here’s what I did instead:

a_number=input(compress(vvalue(any_variable), ' $,'), 16.8);

Let’s unpack that:

1. vvalue takes a variable (numeric or character, doesn’t matter) and outputs the formatted value as character,
2. compress the resulting string to remove spaces, commas, and $s, and
3. input to convert character to numeric.