ODV Output Incorrect Header

[robinskil]

Describe the bug
When the key column is also part of the metadata output for the ODV output, then the header produces another cruise column.

To Reproduce
Specify a key column and re-use that key column also in the metadata as another output column. Then the cruise column is being added twice.

//<Encoding>UTF-8</Encoding>
//<DataField>Ocean</DataField>
//<DataType>Profiles</DataType>
//
//<MetaVariable> label="Cruise" value_type="INDEXED_TEXT" qf_schema="SEADATANET" comment=""</MetaVariable>
//<MetaVariable> label="Station" value_type="INDEXED_TEXT" qf_schema="SEADATANET" comment=""</MetaVariable>
//<MetaVariable> label="Type" value_type="TEXT:2" qf_schema="SEADATANET" comment=""</MetaVariable>
//<MetaVariable> label="Longitude [degrees east]" value_type="FLOAT" qf_schema="SEADATANET" comment=""</MetaVariable>
//<MetaVariable> label="Latitude [degrees north]" value_type="FLOAT" qf_schema="SEADATANET" comment=""</MetaVariable>
//<MetaVariable> label="Cruise" value_type="INDEXED_TEXT" qf_schema="SEADATANET" comment=""</MetaVariable>
//<MetaVariable> label="PLATFORM_NUMBER" value_type="INDEXED_TEXT" qf_schema="SEADATANET" comment=""</MetaVariable>
//<MetaVariable> label="EDMO_CODE" value_type="INDEXED_TEXT" qf_schema="SEADATANET" comment=""</MetaVariable>
//<MetaVariable> label="WMO_INST_TYPE" value_type="INDEXED_TEXT" qf_schema="SEADATANET" comment=""</MetaVariable>
//<MetaVariable> label="BIGRAM" value_type="INDEXED_TEXT" qf_schema="SEADATANET" comment=""</MetaVariable>
//<MetaVariable> label="TIME_QC" value_type="INTEGER" qf_schema="SEADATANET" comment=""</MetaVariable>
//<MetaVariable> label="POSITION_QC" value_type="INTEGER" qf_schema="SEADATANET" comment=""</MetaVariable>
//<MetaVariable> label="PROFILE_QC" value_type="INTEGER" qf_schema="SEADATANET" comment=""</MetaVariable>
//
//<DataVariable> label="DEPTH" value_type="FLOAT" qf_scheme="SEADATANET" comment="" is_primary_variable="T" </DataVariable>
//<DataVariable> label="time_ISO8601" value_type="DOUBLE" qf_scheme="SEADATANET" comment="" is_primary_variable="F" </DataVariable>
//<DataVariable> label="TEMPERATURE" value_type="FLOAT" qf_scheme="SEADATANET" comment="" </DataVariable>
//<DataVariable> label="SALINITY" value_type="FLOAT" qf_scheme="SEADATANET" comment="" </DataVariable>
//
Cruise	Station	Type	yyyy-MM-ddTHH:mm:ss.SSS	Longitude [degrees east]	Latitude [degrees north]	time_ISO8601	DEPTH	QV:SEADATANET:DEPTH	TEMPERATURE	QV:SEADATANET:TEMPERATURE	SALINITY	QV:SEADATANET:SALINITY	DC_REFERENCE	PLATFORM_NUMBER	EDMO_CODE	WMO_INST_TYPE	BIGRAM	TIME_QC	POSITION_QC	PROFILE_QC
10983387	1	*	1974-08-29T06:55:01.000	9.7167	43.95	1974-08-29T06:55:01.000	0.0	1	23.699001	1	38.299004	1	10983387	SHIP	1051	830	CT	1	1	1
10983387	1	*	1974-08-29T06:55:01.000	9.7167	43.95	1974-08-29T06:55:01.000	4.958848998752177	1	23.699001	1	38.289	1	10983387	SHIP	1051	830	CT	1	1	1
10983387	1	*	1974-08-29T06:55:01.000	9.7167	43.95	1974-08-29T06:55:01.000	19.837649708986554	1	23.119001	1	38.209003	1	10983387	SHIP	1051	830	CT	1	1	1
10983387	1	*	1974-08-29T06:55:01.000	9.7167	43.95	1974-08-29T06:55:01.000	24.797009022007295	1	22.469002	1	38.189003	1	10983387	SHIP	1051	830	CT	1	1	1
10983387	1	*	1974-08-29T06:55:01.000	9.7167	43.95	1974-08-29T06:55:01.000	29.75624798295854	1	21.229	1	38.169003	1	10983387	SHIP	1051	830	CT	1	1	1
10983387	1	*	1974-08-29T06:55:01.000	9.7167	43.95	1974-08-29T06:55:01.000	34.71536661041172	1	18.949001	1	38.109	1	10983387	SHIP	1051	830	CT	1	1	1
10983387	1	*	1974-08-29T06:55:01.000	9.7167	43.95	1974-08-29T06:55:01.000	39.67436492293559	1	17.799002	1	38.049004	1	10983387	SHIP	1051	830	CT	1	1	1
10983387	1	*	1974-08-29T06:55:01.000	9.7167	43.95	1974-08-29T06:55:01.000	44.63324293909629	1	16.789001	1	38.019	1	10983387	SHIP	1051	830	CT	1	1	1

[robinskil]

To workaround this issue. Add another instance of the column that is used as the key column to the selected column but with a different alias. Then use that alias as the key column. This will mitigate this issue for the time being.

query_builder = client.query()

query_builder.add_select_column("TIME", alias="TIME")  # Select the columns you want to retrieve
query_builder.add_select_column("TIME_QC", alias="TIME_QC")
query_builder.add_select_column("LONGITUDE") 
query_builder.add_select_column("LATITUDE")
query_builder.add_select_column("POSITION_QC") 
query_builder.add_select_column("DC_REFERENCE")
query_builder.add_select_column("DC_REFERENCE", "ODV_TAG") # Here we re-add as a under a different alias
........others

odv_output = Odv(
    longitude_column=OdvDataColumn("LONGITUDE"),
    latitude_column=OdvDataColumn("LATITUDE"),
    time_column=OdvDataColumn("TIME"),
    depth_column=OdvDataColumn("DEPTH", qf_column="DEPTH_QC"),
    data_columns=[
        OdvDataColumn("TEMPERATURE", qf_column="TEMPERATURE_QC"),
        OdvDataColumn("SALINITY", qf_column="SALINITY_QC"),
    ],
    metadata_columns=[
        OdvDataColumn("DC_REFERENCE"),
    ],
    qf_schema="SEADATANET",
    key_column="ODV_TAG"
)
query_builder.to_odv(odv_output, "output.zip")