# publish.source\_data()

### Parameters

**`source_type`***`:str:`*`Valid values are "csv", "dataframe", or "table"`

**`file_path`***`:str: (Optional)`*`Full path to the csv file to upload. Only pass when source_type="csv"`

**`df`***`:pandas DataFrame: (Optional)`*`DataFrame to upload. Only pass when source_type="dataframe"`

**`table`***`:str: (Optional)`*`Name of existing Snowflake table. Only pass when source_type="table"`

**`data_source_name`***`:str: (Optional)`*`Name for this DataSource`

**`data_source_domain`***`:str: (Optional)`*`Domain for this DataSource`

**`data_source_table_name`***`:str: (Optional)`*`Name to give to the uploaded table in Snowflake. Only pass when source_type is "csv" or "dataframe"`

**`parent_data_source_id`***`:str: (Optional)`*`Parent DataSource for this DataSource`

**`if_exists`***`:bool: (Optional)`*`Instructions on how to proceed when a DataSource exists with this table. Valid values are "fail", "replace", "append".`

{% hint style="info" %}
NOTE: *if\_exists* parameter

"replace" and "append" options are provided to support incremental loading of data through pyrasgo. If you are not operating on an existing DataSource, it is highly recommended that you pass this parameter as "fail". This will warn you if a DataSource exists with this table name.
{% endhint %}

### Return Object

[Rasgo DataSource](https://docs.rasgoml.com/rasgo-docs/rasgo-0.1/pyrasgo-0.3/sources/broken-reference)

### Sample Usage

Upload a csv:

```python
source_type = "csv"
file_path = "Users/me/Downloads/myfile.csv"
data_source_name = "My CSV Test"
data_source_table_name = "CSV_MYFILE_TEST_ONE"
datasource = rasgo.publish.source_data(source_type,
                                       file_path,
                                       data_source_name,
                                       data_source_table_name,
                                       if_exists="replace"
                                       )                                       
print('DataSource:', datasource)
```

Upload a DataFrame:

```python
source_type = "dataframe"
df = myPandasDf
data_source_name = "My DF Test"
data_source_table_name = "DF_MYPANDAS_TEST_ONE"
datasource = rasgo.publish.source_data(source_type,
                                       df,
                                       data_source_name,
                                       data_source_table_name,
                                       if_exists="append"
                                       )                                       
print('DataSource:', datasource)
```

Register a table:

```python
source_type = "table"
table = "SFDATABASE.SFSCHEMA.EXISTING_SF_TBL"
data_source_name = "My Table Test"
datasource = rasgo.publish.source_data(source_type,
                                       table,
                                       data_source_name,
                                       if_exists="fail"
                                       )                                       
print('DataSource:', datasource)
```

### Best Practices / Tips

{% hint style="success" %}
TIP: Operating on existing sources

When **re-loading** data from a csv or DataFrame into an existing source, the \_datasource\_table\_name \_parameter must be included for this function to associate the data with the correct existing table. Omitting this parameter will create a new DataSource.
{% endhint %}