One of the many available transports that can be used to get data into your Data Source, is an existing Attachment Source. Although Attachment Sources are typically used for ancillary project data such as site photos, you can also use them as a staging area for time-series data files, which can then be accessed by one or more Data Sources.
To achieve this when creating a new Data Source, choose "Read from Attachment Source" as the transport type:
And then choose the specific Attachment Source (which must be in the same Workspace):
After this, selecting the data file (or matching multiple files) and configuring the parser is done just like any other Data Source.
Configuring a single Data Source to read from an Attachment Source is a neat trick, but it doesn't really gain you anything; since you still had to send the data file (e.g. via FTP or email) to the Attachment Source to begin with, you could have just as easily sent it directly to the Data Source and saved an intermediate step. However, this technique starts to make a lot more sense when you have multiple Data Sources that need to share the same data file. You only have to send the file once (to the Attachment Source), and then many Data Sources can access that file. This is especially useful when combined with a row-series parser, allowing you to relate specific series IDs in the file to different Data Sources.