Skip to main content
Version: current [25.x]

Categories: Directory

MINDIR

Returns the name of a subdirectory of a table in HDFS. The subdirectory has the name that ranks lowest in case-sensitive alphanumeric order.

Syntax​

MINDIR('dfs', '<path_of_parent_directory>') → <name_of_subdirectory>​

  • dfs: Specifies that the table data is in an HDFS filesystem.
  • <path_of_parent_directory>: The full path of the directory in which the table data is stored.
  • <name_of_subdirectory>: The name of the subdirectory.

Examples

Example 1: Suppose that there is a directory named querylogs in the /tmp directory. In the querylogs directory, there are three subdirectories:

  • 2019
  • 2020
  • 2021

In each subdirectory, there is a CSV file that contains log data. You can use MINDIR to query the data that is in the 2019 subdirectory. The result is the same as if you used the IMINDIR function.

SELECT * FROM dfs."/tmp/querylogs" WHERE dir0 = MINDIR('dfs','/tmp/querylogs')
-- 2019

Example 2: Suppose that there is a directory named products in the /tmp directory. In the products directory, there are these subdirectories in this order:

  • a
  • A
  • B
  • b
  • c
  • C
  • d
  • D

In each subdirectory, there is a CSV file that contains product data. When the query runs, the names of the subdirectories are sorted into case-sensitive order:

  • A
  • B
  • C
  • D
  • a
  • b
  • c
  • d

Because the Unicode values of upper-case letters are lower than the Unicode values of lower-case letters, the lower-case letters follow the upper-case ones. The letter with the lowest Unicode value in this example is A, and that is the name returned in dir0.

Use dir0 to retrieve the name of a subdirectory one level below the parent directory, dir1 to retrieve the name of a subdirectory two levels below, and so on.

SELECT * FROM dfs."/tmp/querylogs" WHERE dir0 = MINDIR('dfs','/tmp/querylogs')
-- A