DataFolder currently includes code for compressing and decompressing multiple multivariate time series while modelardb_compression only supports compressing and decompressing a single univariate time series. Compression of multiple multivariate time series in DataFolder is done by sorting the uncompressed data by tags and timestamps so multivariate time series can be efficiently created through slicing. Afterwards, the univariate time series of each multivariate time series can be compressed and stored as segments. Instead of reimplementing similar functionality in ModelarDB as discussed in #297, maybe just move the code from DataFolder to modelardb_compression. To allow it to be used in different contexts, each part of the pipeline should probably be its own method.
DataFoldercurrently includes code for compressing and decompressing multiple multivariate time series whilemodelardb_compressiononly supports compressing and decompressing a single univariate time series. Compression of multiple multivariate time series inDataFolderis done by sorting the uncompressed data by tags and timestamps so multivariate time series can be efficiently created through slicing. Afterwards, the univariate time series of each multivariate time series can be compressed and stored as segments. Instead of reimplementing similar functionality in ModelarDB as discussed in #297, maybe just move the code fromDataFoldertomodelardb_compression. To allow it to be used in different contexts, each part of the pipeline should probably be its own method.