Stock Screener Data Quality: Missing Metrics, Dates & Currency

Blank means unknown or not applicable

The canonical CSV uses an empty numeric field for missing data. Strings such as N/A, em dashes, percent symbols, abbreviated billions or grouped currency amounts must be mapped deliberately to the documented plain-decimal schema. Zero is a measured zero, not an import default.

Apply compatible definitions

P/E may be trailing, forward, adjusted or unadjusted. Revenue growth can be quarterly year-on-year, annual or forecast. The local tool stores one user-declared comparable definition per column and cannot validate the provider’s accounting or forecast methodology. Do not join incompatible exports without reconciling those definitions.

Use one currency and explicit units

All rows in a dataset must use the selected currency. Price and market capitalization use plain units of that currency; percentage columns use percentage points, so 3 means 3%, not 0.03. There is no FX conversion or conversion from fund assets to equity market capitalization.

Inspect the decision trail

The local tool retains every accepted row. All-source view shows pass/fail reasons, and a decisions CSV preserves the full source decision list. Numeric sorting keeps missing values last in either direction. Repeated symbol labels are flagged and retained; they are not resolved across exchanges or share classes.

Treat timestamps as declarations

Each row has an asOf date. The optional age filter checks0≤referenceDate−asOf≤maximumDays; future-dated and older rows fail it. Without that filter they remain, while future-dated rows are counted in source-quality information. This is a row snapshot check, not a point-in-time audit of every financial input.