Troubleshooting
Configuration validation failed
datacrafter run and datacrafter check call validate_config before work
starts.
- Missing
versionorproject-name - No
extractor/extractorsblock - Unknown extractor, source, or destination
type(the error lists registered names) - Unknown extractor
mode(must besinglefile,api, orcode) - Unknown
methodfor file extractors (url,urlbypattern,apibackuper)
Print the expected schema:
datacrafter config schema
datacrafter config validate
Project configuration file not found
Run commands from a directory that contains datacrafter.yml, or pass
--path:
datacrafter check --path ./my-project
Create a project with datacrafter init if the file is missing.
Environment issues after a valid config
datacrafter check then looks for optional drivers. Install the matching
package (pymongo, python-arango, pycouchdb, meilisearch, pyarrow)
or change the destination type.
APIBackuper extractors need the apibackuper tool on PATH and
storage/apibackuper.cfg in the project.
No JSONL files for schema or metrics
datacrafter schema and metrics
read JSONL under output/ (or current/). Run the pipeline first, or point
--path at a project that already has output.
Extractor downloaded nothing
- Confirm the URL returns the file (TLS is verified by default)
- For
urlbypattern, bothprefixanddata_prefixmust match the HTML index force: truere-downloads even ifcurrent/already has a file
Processor skipped or failed records
error_strategy is skip, fail, or retry. Skipped or failed records go to
output/errors.jsonl. Custom scripts must define process(record) and live
inside the project directory.
ZIP treated as a stream
If the source is a ZIP of XML files, set processor config.type: zipxml and
tagname to the repeating element. A .zip archive is not stream compression
(gz / bz2 / xz / zst).
Secrets in YAML
Use environment interpolation instead of committing credentials:
connstr: "${MONGO_URI}"
# or
connstr: "${MONGO_URI:-mongodb://localhost:27017}"
See Security.