Package org.elasticsearch.xpack.esql.datasources.spi

Class StoragePath

java.lang.Object
org.elasticsearch.xpack.esql.datasources.spi.StoragePath
public final class StoragePath extends Object
Represents a location in a storage system. Uses URI-like format: scheme://[userInfo@]host[:port][/path] Unlike java.net.URI, this class: - Does not perform URL encoding/decoding - Has simpler parsing rules suitable for blob storage keys - Provides convenient methods for path manipulation The userInfo component is preserved verbatim and exposed via userInfo(). Its meaning is provider-defined: for Azure WASB(S) URIs in the canonical Hadoop/Spark form (wasbs://<container>@<account>.blob.core.windows.net/<blob>), the userInfo is the container name. The boundary between userInfo and host is the last '@' in the authority, per RFC 3986. Note: glob pattern detection (isPattern()) only inspects the path component. Glob characters in the scheme or authority (e.g. s3://bucket-*&#47;data/) are not detected as patterns and will be treated as literal text.

  • Field Details

  • Method Details

    • fileUri

      public static String fileUri(Path path)
      Converts a local filesystem Path to a properly formatted file:// URI string. Normalizes Windows backslashes and ensures the URI has three slashes (file:///).

    • ofLocalPath

      public static StoragePath ofLocalPath(Path path)
      Builds a canonical file:// StoragePath from a local filesystem Path. All local producers (directory listings, the query's location, per-object stats keys) must funnel through this so their normalized locations cannot diverge; divergence between two such producers is exactly what broke stats reconciliation on Windows.

    • of

      public static StoragePath of(String location)

    • scheme

      public String scheme()

    • userInfo

      public String userInfo()
      Returns the userInfo component of the URI authority, or null when absent. Empty userInfo (e.g. "scheme://@host/x") is normalized to null. The value is preserved verbatim (no URL decoding) and is provider-defined: Azure WASB(S) URIs in canonical Hadoop form carry the container name here.

    • host

      public String host()

    • port

      public int port()

    • path

      public String path()

    • localPath

      public String localPath()
      Returns the path component adjusted for use as a local filesystem path. On Windows, file:// URIs produce paths like /C:/dir/file where the leading slash is invalid for the OS. This method strips it when a drive letter is detected so the result can be passed to PathUtils.get() safely.

    • objectName

      public String objectName()

    • parentDirectory

      public StoragePath parentDirectory()

    • appendPath

      public StoragePath appendPath(String element)

    • isPattern

      public boolean isPattern()
      Returns true if the path contains glob metacharacters: *, ?, {, [

    • patternPrefix

      public StoragePath patternPrefix()
      Returns a new StoragePath truncated to the longest non-pattern prefix directory. e.g. "s3://b/data/2024/*.parquet" -> "s3://b/data/2024/"

    • globPart

      public String globPart()
      Returns the glob portion of the path (everything after the prefix directory). e.g. "s3://b/data/2024/*.parquet" -> "*.parquet"

    • toString

      public String toString()
      Overrides:
      toString in class Object

    • equals

      public boolean equals(Object o)
      Overrides:
      equals in class Object

    • hashCode

      public int hashCode()
      Overrides:
      hashCode in class Object