Why is the Parquet file format not supported?

The operation succeeds when using streamload curl
**curl -v -L --location-trusted -u "${USER}:${PASSW}" \
-H "format: parquet" \
-H "label: ${LABEL}" \
-H "timeout: ${TIMEOUT}" \
-H "Expect: 100-continue" \
-T "${SOURCE_FILE}" \
"http://${DORIS_HOST}:${DORIS_PORT}/api/${DB_NAME}/${TABLE_NAME}/_stream_load" \
| tee -a "${LOG_FILE}"**

yet an error occurs with streamloader after adding the header format:parquet,
**${DORIS_LOADER} \
--source_file="${SOURCE_FILE}" \
--url="${DORIS_URL}" \
--header="format:parquet" \
--db="${DB_NAME}" \
--table="${TABLE_NAME}" \
--u="${USER}" \
--p="${PASSW}" \
--workers=${WORKERS} \
--check_utf8=false \
2>&1 | tee -a "${LOG_FILE}"**\

as shown below:

regent@dorisFEServer01:~/testdata$ bash streamload_parquet.sh
[2026-01-27 09:00:05] Start Doris Stream Load
time=2026-01-27 09:00:05 level=error msg=Read file failed, error message: EOF, before retrying, we suggest:
1.Check the input data files and fix if there is any problem.
2.Do select count(*) to check whether data is partially loaded.
3.If the data is partially loaded and duplication is unacceptable, consider dropping the table (with caution that all data in the table will be lost) and retry.
4.Otherwise, just retry.
line=reader/reader.go:139
time=2026-01-27 09:00:05 level=error msg=5.When using a specified line delimiter, the file must end with that delimiter. line=reader/reader.go:141

Start with reader/reader.go around lines 139-141 and reproduce the streamloader command using the supplied Parquet header and source-file options. Compare its behavior with the successful curl stream load, then establish whether the loader should support this input and define done as resolving the EOF or clearly reporting the format limitation.

Escrito por el modelo de indexación a partir del texto del issue.