Skip to main content
Importing telemetry is the first step in turning raw data into reliable analysis. Use this page to choose the right import workflow for your file and get the data ready for investigation.
For details on import settings, file type support, and configurations, see Import settings.

​
Before you begin

Complete these steps before starting an import to avoid configuration errors and failed uploads.

​
Choose an import method

MethodBest forRequiresSupported file types
Local file (UI)One-off imports; verify mapping visually before committingBrowser accessCSV, Parquet, TDMS
REST APIRepeatable or automated ingestion pipelinesAPI keyCSV, Parquet, TDMS, Chapter 10
Hosted file URLFiles already in a remote location (HTTP or S3)API keyCSV, Parquet, TDMS, Chapter 10
Python clientScripted ingestion, preprocessing, or engineering tool integrationAPI key; Python environmentCSV, Parquet, TDMS, Chapter 10
If you need an API key, see Create an API key.

​
Gather what you need

Your file must include a timestamp column in one of the supported timestamp formats. Sift cannot import data without a valid timestamp column. This is the most common cause of failed imports. Verify your file has a timestamp column in a supported format before you begin.
Before you start, have answers to these four questions:
  • Which Asset should own this data? An Asset represents the system that generated the data, such as a vehicle or test rig. You can use an existing Asset or create a new one during import.
  • Which Run should this data belong to? A Run is a single data collection session. You can create a new Run or append to an existing one.
  • What is your timestamp format? Sift must know how to interpret the time column in your file. Check your file before you start and match it against the supported timestamp formats.
  • Which columns should become Channels? A Channel is a single time-series signal. You do not need to import every column; only map the ones you intend to analyze.
Imports can be retried. If a mapping is wrong or a timestamp format is misidentified, you can delete the Run and reimport without affecting the Asset or other Runs.
The table below summarizes the file formats supported and the methods available for importing data:

​
Supported formats

FormatSupported in Sift UISupported in sift-stack-py
CSVYesYes
ParquetYesYes
TDMSYesYes
Chapter 10NoYes
For details on import settings, file type support, and configurations, see Import settings.

​
Import from the user interface

  1. In Sift’s homepage, click Import data.
  2. Select your file.
  3. Configure the Asset and Run, time, and Channel configuration settings.
  4. Click Upload.

​
Import using the REST API

  1. Send a configuration request to the CreateDataImportFromUpload endpoint with your API key. The endpoint accepts a format-specific configuration object depending on your file type.
    • Alternatively: To import a file from a remote URL instead of uploading it directly, use the CreateDataImportFromUrl endpoint.
  2. Upload the file using the returned uploadUrl.
  3. Optional: Verify the upload using the returned dataImportId.

​
Import using the Python client

  1. Install the Python client. 
    pip install sift-stack-py
  2. Configure the client with your API key and Sift URL.
  3. Upload the file using the appropriate upload service for your format.