Skip to content

Latest commit

 

History

History
36 lines (33 loc) · 1.76 KB

README.md

File metadata and controls

36 lines (33 loc) · 1.76 KB

Evaluate ModelarDB Changes

This script evaluates what impact a set of changes has on ModelarDB in terms of ingestion time, query processing time, and the amount of storage required. The script automatically computes and evaluates all possible combinations for the set of changes. If a large number of combinations have been evaluated it may be beneficial to analyze the non-dominated set, e.g., computed using pareto.py, instead of the raw output.

The changes must be given in a changes.json file with the following format:

{
  "file/in/modelardb-rs/to/change.rs#first_line_to_remove-last_line_to_remove": {
    "first_line_to_add({})": ["first_input_to_test", "second_input_to_test"],
    "second_line_to_add({})": ["first_input_to_test", "second_input_to_test"]
  }
}

As a concrete example, the following changes.json was used to evaluate the impact of different settings for ArrowWriter for commit 8c0d145:

{
  "server/src/storage/mod.rs#215-216": {
    ".set_encoding(datafusion::parquet::basic::Encoding::{})": ["PLAIN", "PLAIN_DICTIONARY", "RLE", "BIT_PACKED", "DELTA_BINARY_PACKED", "DELTA_LENGTH_BYTE_ARRAY", "DELTA_BYTE_ARRAY", "RLE_DICTIONARY", "BYTE_STREAM_SPLIT"],
    ".set_compression(datafusion::parquet::basic::Compression::{})": ["UNCOMPRESSED", "SNAPPY", "GZIP", "LZO", "BROTLI", "LZ4", "ZSTD"],
    ".set_dictionary_enabled({})": ["false", "true"],
    ".set_statistics_enabled(datafusion::parquet::file::properties::EnabledStatistics::{})": ["None", "Chunk", "Page"]
  }
}