'Does Apache IoTDB's Tsfile have a similar mechanism with Hive's Orc file?
As we know, orc files can store three-level column statistics and stripe in-bloom filters to reduce data out of the library. Does Apache IoTDB's Tsfile have a similar mechanism?
Solution 1:[1]
ORC modeling mode 1 (time, deviceId, S1, s2, S3) stores deviceId repeatedly and also needs to mark whether each point is empty. Modeling mode 3 (time, S1, S2, S3) is not required, but the number of files needs to be considered. TsFile can store time columns for each time series without null values and requires no extra space to mark null values.
Sources
This article follows the attribution requirements of Stack Overflow and is licensed under CC BY-SA 3.0.
Source: Stack Overflow
| Solution | Source |
|---|---|
| Solution 1 | Cecelia Blackt |