View source for Comma-separated values

{{Short description |Text format for tabular data using a comma between fields}}
{{Infobox file format
| name          = Comma-separated values
| icon          = CsvDelimited001.svg
| _noextcode    = on
| extension     = {{code|.csv}}
| _nomimecode   = on
| mime          = <code>text/csv</code><ref name="rfc4180"/>
| type code     = 
| uniform type  = public.comma-separated-values-text<ref name="appleuti">{{cite web |url=https://developer.apple.com/documentation/uniformtypeidentifiers/uttype/3551477-commaseparatedtext |title=commaSeparatedText |work=Apple Developer Documentation: Uniform Type Identifiers |publisher=[[Apple Inc]] |access-date=2023-05-22 |archive-date=2023-05-22 |archive-url=https://web.archive.org/web/20230522150554/https://developer.apple.com/documentation/uniformtypeidentifiers/uttype/3551477-commaseparatedtext |url-status=live }}</ref>
| conforms to   = [[Delimiter-separated values|public.delimited-values-text]]<ref name="appleuti"/>
| magic         = 
| owner         = 
| latest release version = 
| latest release date = 
| genre         = multi-platform, serial data streams
| container for = [[database]] information organized as field separated lists
| contained by  = 
| extended from = 
| extended to   = 
| standard      = RFC 4180
| url           = 
}}

'''Comma-separated values''' ('''CSV''') is a [[plain text]] [[data]] [[content format |format]] for storing [[tabular data]] where the [[field (computer science)|fields]] (values) of a [[record (computer science)|record]] are separated by a [[comma]] and each record is a line (i.e. [[newline]] separated). CSV is commonly-used in [[software]] that generally deals with tabular data such as a [[database]] and a [[spreadsheet]].<ref>{{Cite web |title=Import or export text (.txt or .csv) files |url=https://support.microsoft.com/en-au/office/import-or-export-text-txt-or-csv-files-5250ac4c-663c-47ce-937b-339e391393ba |access-date=2023-08-16 |website=Microsoft Support }}</ref> Benefits cited for using CSV include simplicity of use and [[human-readable |human readability]].<ref>{{Cite web |title=What is a CSV file: A comprehensive guide |url=https://flatfile.com/blog/what-is-a-csv-file-guide-to-uses-and-benefits/ |access-date=2024-10-28 |website=flatfile.com |language=en}}</ref> CSV is a form of [[delimiter-separated values]]. A '''CSV file''' is a file that contains CSV-formatted data.

CSV is not limited to a particular [[character encoding]].<ref name="rfc4180"/> It works just as well with [[Unicode]] (i.e. [[UTF-8]] or [[UTF-16]]) as with [[ASCII]] {{endash}} although particular [[computer program |programs]] may have limitations. Unlike many proprietary data formats, CSV data normally survives naïve translation from one character set to another. CSV does not, however, provide a way to indicate the character encoding, so that must be communicated separately.

== History ==
The CSV format predates [[personal computer]]s by more than a decade. The [[IBM]] [[Fortran]] (level H extended) compiler under [[OS/360]] supported list-directed ("free-form") input/output, with commas between values, in 1972.<ref>{{Citation |title=IBM FORTRAN Program Products for OS and the CMS Component of VM/370 General Information |edition=first |date=July 1972 |id=GC28-6884-0 |url=http://bitsavers.trailing-edge.com/pdf/ibm/370/fortran/GC28-6884-0_IBM_FORTRAN_Program_Products_for_OS_and_CMS_General_Information_Jul72.pdf |page=17 |access-date=February 5, 2016 |quote=For users familiar with the predecessor FORTRAN IV G and H processors, these are the major new language capabilities |archive-date=March 4, 2016 |archive-url=https://web.archive.org/web/20160304113723/http://bitsavers.trailing-edge.com/pdf/ibm/370/fortran/GC28-6884-0_IBM_FORTRAN_Program_Products_for_OS_and_CMS_General_Information_Jul72.pdf |url-status=live }}</ref> List-directed input/output was defined in [[FORTRAN 77]], approved in 1978. List-directed input used commas or spaces for delimiters, so unquoted character strings could not contain commas or spaces.<ref>{{Citation |title=Fortran 77 Language Reference |chapter=List-Directed I/O |chapter-url=http://docs.oracle.com/cd/E19957-01/805-4939/6j4m0vnc5/index.html |publisher=Oracle |access-date=2012-10-26 |archive-date=2021-02-26 |archive-url=https://web.archive.org/web/20210226163504/https://docs.oracle.com/cd/E19957-01/805-4939/6j4m0vnc5/index.html |url-status=live }}</ref>

The term "comma-separated value" and the "CSV" abbreviation were in use by 1983.<ref>{{cite web |url=https://archive.org/stream/Infoworld-1983-09-12#page/n39/mode/2up/search/%22comma-separated+value%22 |title=SuperCalc², spreadsheet package for IBM, CP/M |access-date=December 11, 2017}}</ref> The manual for the [[Osborne Computer Corporation|Osborne]] Executive computer, which bundled the [[SuperCalc]] spreadsheet, documents the CSV quoting convention that allows strings to contain embedded commas.<ref>{{cite web |url=https://archive.org/stream/bitsavers_osborneexeutiveRef1983_16634758/3F00186-00_ExecutiveRef_1983#page/n143/mode/2up/search/%22comma-separated+value%22 |title=Comma-Separated-Value Format File Structure |year=1983 |access-date=December 11, 2017}}</ref>

Comma-separated value lists are easier to type (for example into [[punched card]]s) than fixed-column-aligned data, and they were less prone to producing incorrect results if a value was punched one column off from its intended location.

Comma separated files are used for the interchange of database information between machines of two different architectures. The plain-text character of CSV files largely avoids incompatibilities such as [[endianness|byte-order]] and [[word size]]. The files are largely human-readable, so it is easier to deal with them in the absence of perfect documentation or communication.<ref>{{cite web |url=https://www.loc.gov/preservation/digital/formats/fdd/fdd000323.shtml |title=CSV, Comma Separated Values (RFC 4180) |website=[[Library of Congress]] |access-date=September 22, 2025}}</ref>

The main standardization initiative—transforming "''[[de facto]]'' fuzzy definition" into a more precise and ''[[de jure]]'' one—was in 2005, with {{IETF RFC|4180}}, defining CSV as a [[MIME content type|MIME Content Type]].<ref>{{cite IETF |rfc=4180 |title=Common Format and MIME Type for Comma-Separated Values (CSV) Files |access-date=December 22, 2020}}</ref> Later, in 2013, some of RFC 4180's deficiencies were tackled by a W3C recommendation.<ref>See [http://www.w3.org/TR/sparql11-results-csv-tsv/ sparql11-results-csv-tsv], the first W3C recommendation scoped in CSV and filling some of RFC 4180's deficiencies.</ref>

In 2014 [[IETF]] published {{IETF RFC|7111|link=no}} describing the application of [[URI fragment]]s to CSV documents. RFC 7111 specifies how row, column, and cell ranges can be selected from a CSV document using position indexes.<ref>{{cite IETF |rfc=7111 |title=URI Fragment Identifiers for the text/csv Media Type |access-date=December 22, 2020}}</ref>

In 2015 [[W3C]], in an attempt to enhance CSV with [[Semantic Web|formal semantics]], publicized the first ''drafts of recommendations'' for CSV metadata standards, which began as ''recommendations'' in December of the same year.<ref name="w3c model for tabular data">{{cite web |url=https://www.w3.org/TR/tabular-data-model/ |title=Model for Tabular Data and Metadata on the Web |date=17 December 2015 |access-date=March 23, 2016}} (W3C Recommendation)</ref>

== Specification ==
Casually, CSV refers to data that is [[plain text]] and consists of one record per line where each line has the same sequence of fields separated by a comma.<ref name="rfc4180"/><ref>{{cite web | url=http://edoceo.com/utilitas/csv-file-format | title=Comma Separated Values (CSV) Standard File Format | publisher=Edoceo, Inc | access-date=June 4, 2014 | archive-date=July 14, 2020 | archive-url=https://web.archive.org/web/20200714050729/http://edoceo.com/utilitas/csv-file-format | url-status=live }}</ref><ref>*{{Citation |last= Creativyst |year= 2010 |title= How To: The Comma Separated Value (CSV) File Format |publisher= creativyst.com |access-date= May 24, 2010 |url= http://www.creativyst.com/Doc/Articles/CSV/CSV01.htm |archive-date= April 4, 2021 |archive-url= https://web.archive.org/web/20210404114001/http://www.creativyst.com/Doc/Articles/CSV/CSV01.htm |url-status= live }}</ref> The format is more formally described in the 2005 technical standard {{IETF RFC|4180}} which codifies the CSV format and defines the [[MIME type]] <code>text/csv</code> for the handling of text-based fields. Among its requirements:

* A line is terminated per MS-DOS-style: carriage return and line feed (CR/LF) sequence
* A line terminator is optional for the last line
* The data can start with a header record but with no way to test whether the first line is, in fact, a header, care is required when importing
* Each record should contain the same number of fields
* A field containing a comma, double quote or line terminator character should be enclosed in double quotes
* Any field may be enclosed in double quotes
* If a field is enclosed in double quotes, then a double quote embedded in the field must be represented by a sequence of two double quotes

Common challenges with CSV include:
* Programs may not support line terminator characters within a field even when properly quoted
* Programs may confuse a header line with data or interpret the first data line as a header
* Double quotes in a field may not be parsed correctly

In 2011, [[Open Knowledge Foundation]] (OKF) and various partners created a data protocols working group, which later evolved into the Frictionless Data initiative. One of the main formats they released was the Tabular Data Package. Tabular Data package was heavily based on CSV, using it as the main data transport format and adding basic type and schema metadata. (CSV lacks any type information to distinguish the string <code>1</code> from the number 1.)<ref>{{cite web |title=Tabular Data Package |url=https://frictionlessdata.io/specs/tabular-data-package/ |website=Frictionless Data Specs}}</ref> The Frictionless Data Initiative has also provided a standard CSV Dialect Description Format for describing different dialects of CSV, for example specifying the field separator or quoting rules.<ref>{{cite web |title=CSV Dialect |url=https://frictionlessdata.io/specs/csv-dialect/ |website=Frictionless Data Specs}}</ref>

In 2013, the [[World Wide Web Consortium|W3C]] "CSV on the Web" working group began to specify technologies providing higher interoperability for web applications using CSV or similar formats.<ref>{{cite web|url=http://www.w3.org/2013/csvw/wiki/Main_Page|title=CSV on the Web Working Group|year=2013|publisher=[[World Wide Web Consortium|W3C]] CSV WG|access-date=2015-04-22}}</ref> The working group completed its work in February 2016 and is officially closed in March 2016 with the release of a set of documents and W3C recommendations<ref>{{cite web |url=https://github.com/w3c/csvw |title=CSV on the Web Repository |website=[[GitHub]] }} (on GitHub)</ref> for modeling "Tabular Data",<ref name="w3c model for tabular data" /> and enhancing CSV with [[metadata]] and [[Semantic Web|semantics]]. While the [[well-formed element|well-formedness]] of CSV data can readily be checked, testing validity and canonical form is less well developed, relative to more precise data formats, such as [[XML]] and [[SQL]], which offer richer types and rules-based validation.<ref>{{cite web|url=https://www.csvpath.org/topics/validation/schemas-or-rules|title=Rules Or Schemas|year=2024|publisher=CsvPath Project|access-date=2025-02-13}}</ref>

== Applications ==
CSV is commonly-used for [[data exchange]] and is widely supported by data-oriented [[application software |applications]]. It is often used to move tabular data between programs that natively operate on incompatible data {{endash}} often in formats that are [[proprietary format|proprietary]] or undocumented.<ref name="rfc4180">{{cite IETF |title=Common Format and MIME Type for CSV Files |last=Shafranovich |first=Y. |publisher=[[Internet Engineering Task Force|IETF]] |date=October 2005 |rfc=4180 |page=1 }}</ref><ref>{{cite web|url=https://datahub.io/docs/data-packages/csv|title=CSV - Comma Separated Values|access-date=2017-12-02|archive-date=2021-03-07|archive-url=https://web.archive.org/web/20210307221025/https://datahub.io/docs/data-packages/csv|url-status=dead}}</ref><ref>{{cite web | url=http://www.csvreader.com/csv_format.php | title=CSV Files | access-date=June 4, 2014 | archive-date=April 30, 2021 | archive-url=https://web.archive.org/web/20210430225021/https://www.csvreader.com/csv_format.php | url-status=live }}</ref> A common scenario is moving data from a database to a spreadsheet which, in general, use completely different formats. Most database systems can export as CSV and most spreadsheet programs can import CSV-formatted data; leveraging CSV as an intermediate format. Every major [[E-commerce |ecommerce]] platform provides support for exporting data as a CSV file.<ref>{{Cite web |title=CSV Supported Ecommerce Platforms |url=https://www.rfmcalc.com/supported-platforms/ |access-date=2025-03-09 | website=RFM Calc }}</ref>

CSV is also used for storing data. Common data science tools such as [[Pandas (software)|Pandas]] include the option to export data to CSV for long-term storage.<ref>{{Cite web |title=pandas.DataFrame.to_csv — pandas 2.0.3 documentation |url=https://pandas.pydata.org/docs/reference/api/pandas.DataFrame.to_csv.html |access-date=2023-08-16 |website=pandas.pydata.org}}</ref> Benefits of CSV for data storage include the simplicity of CSV makes parsing and creating CSV files easy to implement and fast compared to other data formats, human readability making editing or fixing data simpler,<ref>{{Cite web |date=2021-09-15 |title=CSV Format: History, Advantages and Why It Is Still Popular |url=https://bytescout.com/blog/csv-format-history-advantages.html |access-date=2023-08-16 |website=ByteScout |language=en-US}}</ref> and high compressibility leading to smaller data files.<ref>{{Cite web |date=2020-07-23 |title=Comparison of different file formats in Big Data |url=https://www.adaltas.com/en/2020/07/23/benchmark-study-of-different-file-format/ |access-date=2023-08-16 |website=www.adaltas.com |language=en}}</ref> Alternatively, CSV does not support more complex data relations and makes no distinction between null and empty values, and in applications where these features are needed other formats are preferred.

More than 200 local, regional, and national data portals, such as those of the [[UK government]] and the [[European Commission]], use CSV files with standardized [[Database catalog|data catalogs]].<ref>{{Cite conference |last1=Mahmud |first1=S M Hasan |last2=Hossin |first2=Md Altab |last3=Jahan |first3=Hosney |last4=Noori |first4=Sheak Rashed Haider |last5=Bhuiyan |first5=Touhid |title=CSV-ANNOTATE: Generate annotated tables from CSV file |conference=2018 International Conference on Artificial Intelligence and Big Data (ICAIBD) |publisher=IEEE |publication-date=2018 |pages=71–75 |doi=10.1109/ICAIBD.2018.8396169 |isbn=978-1-5386-6987-7}}</ref>

Some applications use CSV as a [[Data exchange|data interchange format]] to enhance its [[interoperability]], exporting and importing CSV. Others use CSV as an internal format. CSV is supported by almost all spreadsheets and database management systems.

[[Spreadsheet]]s including Apple [[Numbers (spreadsheet)|Numbers]], [[LibreOffice Calc]], and [[Apache OpenOffice]] Calc. support reading CSV files. [[Microsoft Excel]] also supports a dialect of CSV with restrictions in comparison to other spreadsheet software (e.g., {{as of|lc=y|2019}} Excel still cannot export CSV files in the commonly used UTF-8 character encoding, and separator is not enforced to be the comma). [[LibreOffice Calc]] CSV importer is actually a more generic delimited text importer, supporting multiple separators at the same time as well as field trimming.

Various [[relational databases]] support saving query results to a CSV file. [[PostgreSQL]] provides the <code>COPY</code> command, which allows for both saving and loading data to and from a file. {{code|2=postgres|COPY (SELECT * FROM articles) TO '/home/wikipedia/file.csv' (FORMAT csv)}} saves the content of a table <code>articles</code> to a file called <code>/home/wikipedia/file.csv</code>.<ref>{{cite web|url=https://www.postgresql.org/docs/current/sql-copy.html |title=Documentation: 14: COPY |publisher=PostgreSQL |accessdate=2024-05-12}}</ref> Some relational databases, when using standard SQL, offer ''foreign-data wrapper'' (FDW). For example, PostgreSQL offers the {{code|2=postgres|CREATE FOREIGN TABLE}}<ref>{{cite web|url=https://www.postgresql.org/docs/current/postgres-fdw.html |title=Documentation: 14: F.35. postgres_fdw |publisher=PostgreSQL |date=2022-02-10 |accessdate=2022-03-04}}</ref> and {{code|2=postgres|CREATE EXTENSION file_fdw}}<ref>{{cite web|url=https://www.postgresql.org/docs/current/file-fdw.html |title=Documentation: 14: F.14. file_fdw |publisher=PostgreSQL |date=2022-02-10 |accessdate=2022-03-04}}</ref> commands to configure any variant of CSV. Databases like [[Apache Hive]] offer the option to express CSV or [[gzip|.csv.gz]] as an internal table format.

Programs that work with CSV may have limits on the maximum number of rows CSV files can have. Examples include Microsoft Excel (1,048,576 rows), Apple Numbers (1,000,000 rows), Google Sheets (10,000,000 cells), and OpenOffice and LibreOffice (1,048,576 rows).<ref>{{cite web | url=https://www.import2.com/csv/understanding-csv-and-row-limits | title=Understanding CSV and row limits | access-date=Feb 28, 2021 | archive-date=January 15, 2021 | archive-url=https://web.archive.org/web/20210115142749/https://www.import2.com/csv/understanding-csv-and-row-limits | url-status=dead }}</ref>

== See also ==

* {{anl |Comparison of data-serialization formats}}
* [[Delimiter collision]] {{endash}} Formatting challenge caused by data that contains a delimiter
* {{anl |Flat-file database}}
* {{anl |Tab-separated values}}

== References ==

{{refs}}

== Further reading ==

* {{cite web |title=IBM DB2 Administration Guide - LOAD, IMPORT, and EXPORT File Formats |publisher=[[IBM]] |url=https://www.columbia.edu/sec/acis/db2/db2d0/db2d053.htm |access-date=2016-12-12 |url-status=live |archive-url=https://web.archive.org/web/20161213014111/https://www.columbia.edu/sec/acis/db2/db2d0/db2d053.htm |archive-date=2016-12-13}} (Has file descriptions of delimited ASCII (.DEL) (including comma- and semicolon-separated) and non-delimited ASCII (.ASC) files for data transfer.)

{{Data Exchange}}

[[Category:Delimiter-separated formats]]
[[Category:Open formats]]
[[Category:Spreadsheet file formats]]