std/csv

Index

• type CsvErr
  • method CsvErr.display
• type Reader
  • method Reader.read
• fn parse
• fn reader
• fn parse_maps
• fn format
• fn valid_sep
• fn valid_comment

Types

CsvErr

struct CsvErr {
  start_line: Int
  line: Int
  column: Int
  msg: Str
}

Where and why the text is not valid CSV (Go’s csv.ParseError): the line the record starts on, the line and 1-based byte column of the error, and Go’s message (bare " in non-quoted-field, extraneous or missing " in quoted-field, wrong number of fields). Displays as Go’s Error(): parse error on line 1, column 5: bare " in non-quoted-field, record on line 2; parse error on line 3, column 1: ..., record on line 4: wrong number of fields.

Source: std/csv.xo:14

CsvErr.display

fn (e: CsvErr) display() -> Str

Source: std/csv.xo:21

Reader

struct Reader {}

Records read one at a time from an open file (Go’s csv.NewReader(file), Read), the options as parse’s: the file is read as the records are asked for, so a file larger than memory can be walked. A quoted field may hold line breaks.

Source: std/csv.xo:33

Reader.read

fn (var r: Reader) read() -> List[Str]? ! CsvErr uses fs.read

The next record, None at the end of the file. A parse error is as parse’s; a failure of the file is a CsvErr with line 0 and the file’s error text as msg.

Source: std/csv.xo:39

Functions

parse

fn parse(text: Str, sep: Rune = ',', comment: Rune = '\0', fields: Int = -1, lazy_quotes: Bool = false, trim_leading_space: Bool = false) -> List[List[Str]] ! CsvErr

The records of text, as Go’s csv.Reader reads them (decision 0298): sep separates fields (Comma), lines starting with comment are skipped (Comment; '\0', the default, for none), fields is FieldsPerRecord (-1 any number, the default here; 0 the first record’s; N exactly N), lazy_quotes and trim_leading_space are LazyQuotes and TrimLeadingSpace.

Source: std/csv.xo:28

reader

fn reader(f: File, sep: Rune = ',', comment: Rune = '\0', fields: Int = -1, lazy_quotes: Bool = false, trim_leading_space: Bool = false) -> Reader

Source: std/csv.xo:35

parse_maps

fn parse_maps(text: Str, sep: Rune = ',', comment: Rune = '\0', lazy_quotes: Bool = false, trim_leading_space: Bool = false) -> List[Map[Str, Str]] ! CsvErr

The records after the first, each a map from the first record’s fields (the header) to its own; every record must have the header’s number of fields (wrong number of fields).

Source: std/csv.xo:43

format

fn format(rows: List[List[Str]], sep: Rune = ',', crlf: Bool = false) -> Str

The text of rows as Go’s csv.Writer writes them (crlf: UseCRLF): a field is quoted when it holds sep, a quote, \r, or \n, starts with a space, or is \.; also when it is the only, empty field of a record (so it reads back as a record, where Go writes an empty line).

Source: std/csv.xo:48

valid_sep

fn valid_sep(r: Rune) -> Bool

Go’s validDelim (reader.go): whether r can be the sep of parse, parse_maps and format. False for '\0', ", \r, \n and U+FFFD, and for a code point that is not a Unicode scalar value (which no Rune holds: Rune(n) faults on it); every other code point works, a multi byte one included.

Source: std/csv.xo:54

valid_comment

fn valid_comment(r: Rune) -> Bool

Go’s validDelim for comment, and true for '\0', Go’s zero value and the default (no comment). A comment must also differ from sep: parse, parse_maps and format fault csv: invalid field or comment delimiter when it does not.

Source: std/csv.xo:59