Storage Formats

StockSharp supports two market data storage formats defined by the StorageFormats enumeration: Binary and CSV. Each format has its own advantages and is suitable for different use cases.

Binary Format

The binary format is the primary high-performance data storage format in StockSharp. The implementation is located in the BinaryMarketDataSerializer class within Algo/Storages/Binary/.

Features

• Compactness -- data is serialized in binary form with compression, ensuring minimal file sizes.
• Performance -- reading and writing are significantly faster compared to text formats.
• Metadata -- the BinaryMetaInfo class stores auxiliary information: first and last prices, fractional values, timestamps. This allows quickly retrieving general data information without fully reading the file.
• Compression-ready architecture -- the format is designed from the ground up for efficient streaming compression.

Binary files have the .bin extension.

Supported Data Types

The binary format supports serialization of all major market data types: candles, ticks (trades), order books (Level2), Level1 data, orders, and own trades. Each type has a specialized serializer optimized for the structure of the specific data.

CSV Format

The CSV (Comma-Separated Values) text format is implemented in the CsvMarketDataSerializer class located in Algo/Storages/Csv/.

Features

• Readability -- files can be opened in any text editor or in Excel for visual data analysis.
• Editability -- data can be manually corrected when necessary.
• Metadata -- the CsvMetaInfo class provides metadata storage with encoding support. It contains the IncrementalOnly property (incremental data support) and LastId (last record identifier).
• Larger file size -- text representation takes significantly more disk space.
• Slower processing -- parsing text data requires additional computational resources.

CSV files have the .csv extension.

When to Use Which Format

File Organization on Disk

Data on disk is organized according to the following path structure:

{root_folder}/{first_letter}/{instrument_identifier}/{yyyy_MM_dd}/{file_name}.{extension}

Where the extension is .bin for binary format or .csv for text format. This hierarchical organization ensures fast data lookup by instrument and date.

For example, 5-minute candles for the SBER@TQBR instrument on April 1, 2024 in binary format would be located at a path like:

Storage/S/SBER@TQBR/2024_04_01/candles_5m.bin

Converting Between Formats

StockSharp allows loading data in one format and saving it in another. This can be useful, for example, for exporting binary data to CSV for analysis in external tools:

// Load from binary storage
var binaryStorage = storageRegistry.GetTimeFrameCandleMessageStorage(
    securityId, TimeSpan.FromMinutes(5), StorageFormats.Binary);
var candles = await binaryStorage.LoadAsync(from, to).ToArrayAsync();

// Save to CSV storage
var csvStorage = storageRegistry.GetTimeFrameCandleMessageStorage(
    securityId, TimeSpan.FromMinutes(5), StorageFormats.Csv);
await csvStorage.SaveAsync(candles);

Code Example

The format is selected when creating a storage via StorageRegistry:

var storageRegistry = new StorageRegistry();

// Create candle storage in binary format
var binaryStorage = storageRegistry.GetTimeFrameCandleMessageStorage(
    securityId,
    TimeSpan.FromMinutes(5),
    StorageFormats.Binary);

// Create candle storage in CSV format
var csvStorage = storageRegistry.GetTimeFrameCandleMessageStorage(
    securityId,
    TimeSpan.FromMinutes(5),
    StorageFormats.Csv);

// Load data
var from = new DateTime(2024, 1, 1);
var to = new DateTime(2024, 1, 31);
var candles = await binaryStorage.LoadAsync(from, to).ToArrayAsync();

The storage format can also be specified when working with tick data and order books:

// Ticks in binary format
var tickStorage = storageRegistry.GetTickMessageStorage(
    securityId, StorageFormats.Binary);

// Order books in CSV format
var depthStorage = storageRegistry.GetQuoteMessageStorage(
    securityId, StorageFormats.Csv);

See Also

• Working with the API
• Storage Drives