Checking the files - Amazon Lookout for Equipment

Checking the files

If Lookout for Equipment fails to ingest a particular file, consider the following possibilities:

  • None of the sensors listed in the file have any data that can be ingested.

  • The file is not a .csv file, or the file is corrupted, or the file cannot be read for some other reason.

To troubleshoot files that were not ingested:

  1. From the Job details tab of the main console page for your dataset, note the names of any files that failed the ingestion process.

  2. To address issues with file formatting, see Formatting your data.

  3. To address issues with individual sensors, see Understanding sensor quality.

  4. When you’re ready to try again, see Replacing your dataset.

Important

This page is about troubleshooting the ingestion of specific files. You can also read about why the ingestion of an entire job can fail, and about evaluating the data from specific sensors.

Anticipating schema detection problems

The following circumstances will lead to the failure of an entire ingestion job:

  • One or more column headers contain one or more invalid characters.

    A single invalid character in a single column in a single file is enough to fail an entire job involving multiple files.

  • In a job consisting of a single file, that file has a formatting issue that prevents ingestion.

  • In a job consisting of multiple files, every single file has a formatting issue that prevents ingestion.

The easiest way to prevent problems with file ingestion is to take the following precautions:

  • Make sure your headers don't include any invalid characters, such as spaces.

    Valid characters are: 0-9, a-z, A-Z, and # $ . \ - (hyphen) _ (underscore)

  • Make sure that the timestamp column is the one furthest to the left in your CSV file.

  • Make sure that you don't have any duplicated column headers.