Currently two types of data validation tests are available, which are run manually via the Makefile as documented in Import-Process:

$ make test-input

and

$ make test-output

These run the corresponding R files in /tests/test_input and /tests/test_output, respectively. More specifically:

Input tests

Tests of the delivery_metadata.txt file:

• Test that all csvs in original_data are mentioned in dengue_case_metadata.txt.
• Test that dengue_case_metadata.txt doesn't have orphan entries.
• Test that delivery_year is correct for entries in metadata.

Tests for file-specific codes for deliveries:

• Test that all deliveries have file specific codes (does not check for orphan codes).

Output tests

Tests for deliveries:

• Test that unique deliveries are same in all the tables.

Tests of the delivery_metadata table:

• Test that the delivery_metadata table matches data from dengue_case_metadata.txt.
• Check for date_delivered existing.

Tests for expected tables being created in various steps:

• Check if a given table identifier is present in the database.
• Test that the table delivery_metadata is present.
• Test that tables are getting created as a result of running 2_create_unique_data.R.