Skip to content

Latest commit

 

History

History
71 lines (68 loc) · 3.45 KB

omop_52_csv.md

File metadata and controls

71 lines (68 loc) · 3.45 KB
Raw

OMOP CDM v5.2 CSV Datasets

When you specify a dataset_type of omop:5.2:csv in your configuration, rex_deliver_dataset will allow a set of CSV files structured according to v5.2 of the OMOP Common Data Model. The requirements for this type of dataset are as follows:

  • Only data for the tables defined in the following sections of the OMOP specification are allowed:
    • Clinical Data Tables
    • Health System Data Tables
    • Health Economics Data Tables
    • Derived Elements
    • Metadata
  • Data for the tables defined in the Vocabulary section of the OMOP specification is not allowed.
    • The requirement is that data in tables from the other areas of the CDM that refer to records in the Vocabulary tables will be using the standard content (e.g., as published by Athena).
  • The data for each table in OMOP CDM must be delivered as a separate file.
    • The base names of the files must be exactly as the tables are named in OMOP (including underscores, if used).
    • The names can be in any case. E.g., the file that contains data for the PERSON table could be named PERSON.csv, person.CSV, or PeRsOn.CsV.
    • All files must have an extension of .csv.
    • There can only be one file delivered per table. You cannot provide both a PERSON.csv and a person.csv.
  • Each file must contain all columns defined for the given OMOP table, even if they’re not being used.
    • The column names must be listed as the first record in the file.
    • Column names are case-insensitive.
    • The ordering of the columns in the file is not defined. You may provide them in any order, as they all exist.
    • No columns beyond those specified by OMOP can be present in the file.
  • The contents of the files must be structured as Comma-Separated Values files as described in section 2 of RFC4180. Notably:
    • Each record is delimited by a CRLF (ASCII 13 & 10).
    • Each column is delimited by a comma (ASCII 44).
    • Column values may be enclosed in double quotes (ASCII 34).
    • Column values that contain commas, double quotes, or CRLFs must be enclosed in double quotes.
    • All records must have the same number of columns.
  • Column values must be appropriately formatted according to the types specified by OMOP:
    • integer
      • Must be represented as whole numbers using Arabic numerals (ASCII 48 through 57)
      • Commas or other digit grouping separators are not permitted.
    • float
      • Must be represented as decimal numbers using Arabic numerals (ASCII 48 through 57)
      • A period (ASCII 46) must be used to separate the whole number from the fractional part.
    • varchar/text/clob/string
      • Can be any string that fits within the length restrictions specified by OMOP
    • date
      • Must be represented as YYYY-MM-DD (e.g., 2019-05-22)
    • datetime
      • Must be represented as YYYY-MM-DDTHH:MM:SS (e.g., 2019-05-22T12:34:56)
      • If a timezone is being specified, it must be appended to the date as ±HH:MM (e.g. 2019-05-22T12:34:56+04:00)
      • If the timezone should be interpreted as UTC, then either no timezone offset should be specified, or use the single letter Z in the place of the offset (e.g. 2019-05-22T12:34:56Z)
      • Fractional seconds are not allowed.
  • Columns defined as required in the OMOP specification must have values provided in every record.