Investigation: changing data type for Specimen.percent_tumour_cells

[Buwujiu]

The Problem

As part of the clinical exceptions requests, we want to add two text values to Specimen schema's percent_tumour_cells. The new values are:

• Unknown
• Not applicable

See complete requirement here: https://wiki.oicr.on.ca/display/icgcargotech/Core+Clinical+Data+Exceptions#CoreClinicalDataExceptions-UserStories

Currently in argo dictionary, percent_tumour_cells only accepts numbers: https://docs.icgc-argo.org/dictionary

Changing the data type will affect downstream services, for instance the clinical index mapping, donor aggregator, lectern, clinical service.

The solution

Investigate what will be affected by the data type change. Questions to start with:

• how to accept multiple types?
• how to express number + Unknown + NA in lectern?
• what's the best practice in ES for such data type change?
• maybe creating our own type: number + unknown + NA?

Exit Criteria:

• document the outcome of the research

[Buwujiu]

there is a potential change to percent_tumour_cells data type to change to text. If this goes through, there is no need for exception. Talk to @joneubank about potential data migration.

[Buwujiu]

@hknahal to investigate if we can add 2 more fields to describe percent_tumour_cell range. Instead of changing the data type.