Parquet files are used to store the data from log
data files.
You can find more information about the Apache Parquet file format here.
A log
data file will have one or more parquet files associated with it. For each CAN message listed in the log
data file, there will be a corresponding parquet file.
In this example, CAN message MESSAGE_100
has hex ID 0100
and CAN message MESSAGE_101
has hex ID 0101
.
For this log
file, there would be two parquet files:
MESSAGE_100.parquet
with 2 rowsMESSAGE_101.parquet
with 3 rows
The parquet files contain the following columns:
mid
- Corresponds to the "Message Number" column in thelog
file.tm
- The timestamp at which the message was recorded.bus
- Corresponds to the "Bus" column in thelog
file.type
- Corresponds to the "Type" column in thelog
file.- Remaining columns - One for each CAN signal in the CAN message.
- The values stored in the parquet files are not the actual values. They must have the CAN factor and offset applied in order to calculate the actual values.
Users with at least OEM access can download the parquet files for a log
data file. This is useful for debugging the raw data associated with a data file.
- Go to the data files page.
- Select a
log
data file. - Click the
button in the preview overlay.
The resulting ZIP file will contain the parquet files associated with the given data file. If there were any errors accessing any of the parquet files, their file names will be listed in a missing-parquets.txt
file in the ZIP file.