File structure of data export

    Each directory in a database corresponds to a directory in the file structure. The directory hierarchy in the file structure corresponds the directory hierarchy in the database. If some DB directory contains no objects (neither tables nor subdirectories), this directory’s file structure contains a single zero size file named .

    • The scheme.pb file with information about the table structure and its parameters in format.
    • One or more data_XX.csv files with data in CSV format, where XX is the file sequence number. Data export starts with the file. Each subsequent file is created once the size of the current file exceeds 100 MB.

    Data is stored in .csv files, one file line per table entry, without a row with column headers. URL-encoded format is used for string representation. For example, a file line for a table with uint64 and uft8 columns containing the number 1 and the string “Hello”, respectively, looks like this:

    When exporting tables created within a tutorial when Getting started with YQL in the “Getting started” section, the following file structure is created:

    1. ├── episodes
    2. │   └── scheme.pb
    4. │   ├── data_00.csv
    5. │   └── scheme.pb
    6. └── series