Troubleshooting

This section of the documentation contains a couple of common errors encountered while implementing nf-schema or using pipelines with the plugin and how to solve them.

An integer value is failing because a string value was expected

Pipelines often use some kind of identifier field in their sample sheet, that usually gets string as the expected input. This can cause some issues when using CSV and TSV sample sheets since the types of all values are automatically inferred for these file types. Meaning that a value in the sample field that has 123 as the value would result in an integer instead of a string. Following error would be shown in this case:

-> Entry 1: Error for field 'sample' (123): Value is [integer] but should be [string]

There are multiple ways to solve this issue as pipeline developers and as pipeline users. The solutions will be listed from best options to worst:

Pipeline user

• Option 1: Use a YAML or JSON file as sample sheet if that is allowed by the pipeline and make sure to quote the value. ('123' in the case of the example above)
• Option 2: Activate lenient validation mode. Keep in mind that this will enable lenient validation for ALL validations in the pipeline.

Pipeline developer

• Option 1: Modify "type": "string" in the affected JSON schema to "type": ["string", "integer"]. Make sure that integer values don't break anything in your pipeline. The easiest way to ensure compatibility is to cast the value to a string after converting the sample sheet to a list.
• Option 2: Enforce usage of JSON and YAML sample sheets by only allowing files ending in .yml, .yaml or .json to be used.