Troubleshooting/Common Questions
Problems sometimes arise when attempting to create a Google Fusion Tables workspace. This is almost always due to a misconfiguration in the user’s environment. The following suggestions can often help detect and overcome such problems.
My workspace no longer runs. Why don't I have to log in?
Both spatial and non-spatial versions of this format previously used the Google Fusion Tables SQL API to access data using a Google account e-mail and password. However, as of January 2013, that API was deprecated and the new API is incompatible due to changes in the authentication scheme.
As of FME 2015, both versions of the reader and writer use the Google Fusion Tables v1.0 using OAuth 2.0 authentication. This means that an e-mail address and password are no longer required: to authenticate, you will need to obtain authentication via the Named Connection parameter.
To create a valid Named Connection: Click the browse button next to the Refresh Token field in the reader or writer parameters dialog. See the reader overview and reader parameters for more information.
Note: You will need to re-add the reader or writer to older workspaces so they will run properly.
Why did my location not produce geometry?
The reader will only create geometry from data inside Location type fields. The two types of data that the reader currently handles are:
- a comma-/space-separated longitude/latitude pair (for example, -123.1207375,49.2827291)
- KML geometry
What happens if my location column contains different types of data? (for example, KML, lat/lon and geocoded addresses)
The reader will read and create geometry from the types of data it supports (KML and lat/lon pairs) while ignoring other types of data.
Location columns in my table contain unsupported location types. How will I extract that information?
All fields will be attributes on the Fusion table’s feature type. It is up to the user to transform this data using FME’s built in functions. If your table does not have any supported location types, consider using the non-spatial version.
Why are column names sometimes appended with an integer?
Google Fusion tables allows for duplicate column names. In FME, attributes must have unique names. If a duplicate column is read, the first will retain the original name while all following duplicates will be appended with an integer starting from 0.
How do I access public tables?
Inside the Reader Parameters box, add (space-delimited) the Table ID of the public table in the table list text line. The feature type names of public tables will be their table ID. For writing, specify the writer feature type name as the table ID in the Writer Parameters box.
Note that, due to volume, the list of public tables is not available from the API, so explicit IDs are required. Numeric IDs are also no longer supported, so please use the encrypted ID. Both explicit IDs and encrypted IDs are available from the File > About menu for a table when viewed in the Google Fusion Table web interface.
What is the difference between Google Fusion Tables (Spatial) and Google Fusion Tables (Non-spatial) formats?
The non-spatial version will only read attribute data, it is up to the user to transform the data. The spatial reader does everything the non-spatial version does but will also create geometry from KML geometry or latitude/longitude pairs residing under a single location column. This may be convenient for some users with such information in a table.
I am getting the error Unable to create a connection to the Google Fusion Tables server. Check your login information, table names and table IDs. Is this an FME problem or is the service down?
If you are still getting this error after examining your login, table information, network connection, then it is probably an issue with the service.
I am getting an error about geometry columns not having the correct coordinate system. How can I fix this?
Google Fusion Tables writes geometry as KML in the LL84 projection, and requires input features to have a valid coordinate system (ideally LL84) to write or reproject correctly. Using transformers such as the CoordinateSystemSetter may help resolve this issue by identifying the coordinate system of the input features if known.
How can I write to blank or unnamed columns?
The Google Fusion Tables writer needs a unique column name to write correctly. Providing names to the columns of existing tables can be performed through the Google Fusion Tables web interface under Edit > Modify Columns. Note that column names with spaces are currently not supported.
How can I write a single geometry column to a table containing multiple location columns?
The Google Fusion Tables recognizes when multiple geometries exist on input features or multiple location columns exist in the table and will attempt to match geometry names to column names to map them correctly. Without a name, it is not known which column to write to. One solution is to use a GeometryPropertySetter transformer to set a name on the single input geometry that will match the intended destination column name in the table.
Note that when dealing strictly with single geometry on the input features and the table columns, no matching takes place and writing occurs regardless of names.