Structure Options

The optional [structure] section stores a small allowlist of CSV reader options inside an INC file. These options describe how to parse the CSV component after the metadata block.

[structure] is intentionally limited. It is not a place to store every CSV.jl keyword argument. Options outside the allowlist can still be passed directly to readinc as Julia keyword arguments.

Explicit keyword arguments passed to readinc always override values from [structure].

Allowed Keys

Key	Type	Meaning
`delim`	character	Delimiter between CSV fields.
`delimiter`	character	Alias for `delim`; if both appear, `delimiter` wins.
`quotechar`	character	Quote character used by the CSV component.
`escapechar`	character	Escape character used by the CSV component.
`comment`	string	Comment marker used by the CSV component.
`header`	integer	CSV component line containing column names.
`footerskip`	integer	Number of trailing CSV component rows to ignore.

The key names are based primarily on CSV.jl keyword arguments. The alias delimiter is accepted for interoperability with other INC implementations.

Value Rules

Values in [structure] use the same metadata value rules as the rest of the INC metadata block:

unquoted signed integers are read as Int;
quoted values are read as strings;
unquoted non-integer values are read as strings.

Character options accept:

a one-character string, such as delim = "|" or quotechar = "'";
tab for the tab character;
space for a space character;
\t for the tab character;
an integer Unicode code point, such as delim = 44 for comma.

Integer options such as header and footerskip must be unquoted integers. For example, use header = 2, not header = "2".

Line-oriented options are relative to the CSV component after the closing INC metadata delimiter. They are not counted from the physical first line of the INC file.

Examples

Semicolon-Delimited Data

---
title = Semicolon data
[structure]
delim = ";"
---
name;score
Ada;21
Babbage;12

using DataFrames
using IncCSV

file = readinc("semicolon.inc", DataFrame)
table(file)

The alias delimiter can also be used:

[structure]
delimiter = ;

If both delim and delimiter appear, delimiter takes precedence.

Tab-Delimited Data

---
title = Tab data
[structure]
delim = tab
---
name	score
Ada	21
Babbage	12

file = readinc("tab.inc", DataFrame)

Commented CSV Rows

comment applies only to the CSV component. Metadata comments are always # or ;.

---
title = Commented data
[structure]
comment = "#"
---
name,score
# ignored row
Ada,21

file = readinc("commented.inc", DataFrame)

Non-Default Quote Character

---
title = Quote character data
[structure]
quotechar = "'"
---
name,note
Ada,'hello, world'

file = readinc("quotechar.inc", DataFrame)
table(file).note

Non-Default Escape Character

---
title = Escape character data
[structure]
escapechar = |
---
name,note
Ada,"say |"hi|""

file = readinc("escapechar.inc", DataFrame)
table(file).note

Header On A Later CSV Line

header is counted from the first line of the CSV component, after the closing metadata delimiter.

---
title = Later header
[structure]
header = 2
---
discard,discard
name,score
Ada,21

file = readinc("later_header.inc", DataFrame)

---
title = Footer data
[structure]
footerskip = 1
---
name,score
Ada,21
Babbage,12
TOTAL,33

file = readinc("footer.inc", DataFrame)

Explicit Keyword Arguments

Caller-supplied keyword arguments override [structure] values:

file = readinc("data.inc", DataFrame; delim='|')

Use explicit keyword arguments for CSV.jl options outside the [structure] allowlist, such as skipto, limit, missingstring, dateformat, normalizenames, ignoreemptyrows, ignorerepeated, decimal, or groupmark.

writeinc does not infer or create [structure] metadata from CSV writing keyword arguments. If a file needs structure metadata, provide it explicitly in the metadata dictionary.