The PocketCsvReader.Compression namespace provides DecompressorFactory and multiple classes implementing the IDecompressor interface. These classes support decompressing files using GZip and Deflate formats. The class handling Zip supports deflate compression, but also accounts for the fact that a ZIP archive may contain multiple files.
Specifying the Compression to the Reader
You can specify compression using the ResourceDescriptorBuilder class within a CsvProfile. The WithCompression method accepts a compression name either as a string or as a CompressionFormat enumeration value. Alternatively, call WithoutCompression to indicate that the file is not compressed — this is also the default.
var profile = new CsvProfile(
new DialectDescriptorBuilder()
.WithDelimiter(',')
.WithLineTerminator("\n")
.Build(),
null,
new ResourceDescriptorBuilder()
.WithEncoding("utf-8")
.WithCompression("gz")
.Build());
When using compression, it’s strongly recommended to specify the encoding of the underlying file. This significantly improves memory efficiency. See below for more context.
Supported Compressions
The library supports the following formats:
Deflate, with aliaseszzanddef: RFC 1951GZip, with aliasgz: RFC 1952-
Zip, with aliaszipfile GetSupportedKeys()returns all known aliases.GetSupportedCompressions()returns the three supported formats.GetCompression(string alias)returns the primary name for a given alias.
Internal Classes
Class DecompressorFactory
The DecompressorFactory offers two instantiation options: Streaming and Buffered. For better memory performance, prefer Streaming, but note that it’s forward-only, meaning you cannot detect encoding. If encoding cannot be predefined, use the Buffered version instead.
Single File vs Archive
Formats like GZip and Deflate compress one stream to another — effectively one file into one file. In contrast, an archive contains metadata and multiple file streams.
When using an archive (e.g., with ZipDecompressor), the implementation assumes the archive contains a single file. If it doesn’t, decompression fails. This constraint exists because the library is focused on delimited (CSV-like) files only.
