CSV

Parsing

Serde.SerdeCsv.parse_csv — Function

parse_csv(x::Union{AbstractString, Vector{UInt8}}; delimiter = ",", kw...) -> Vector{NamedTuple}

Parse a CSV string or byte vector into a vector of NamedTuples without mapping to a target type.

All values are returned as String; conversion to numeric types is performed later by from_csv / Serde.deser.

Arguments

x: CSV text as a String or Vector{UInt8}.

Keyword arguments

delimiter::AbstractString = ",": field separator.
Additional keyword arguments are forwarded to CSV.File.

Returns

A Vector of NamedTuples (one per data row), with keys from the header row.

Throws

ParseError: if x is not valid CSV.

Examples

julia> parse_csv("name,age\nAlice,30\nBob,25")
2-element Vector{NamedTuple}:
 (name = "Alice", age = "30")
 (name = "Bob",   age = "25")

Deserialization

Serde.SerdeCsv.from_csv — Function

from_csv(::Type{T}, x; kw...) -> Vector{T}
from_csv(strategy, ::Type{T}, x; kw...) -> Vector{T}
from_csv(f::Function, x; kw...) -> Any

Parse a CSV string and deserialize each row into type T, returning a Vector{T}.

Each CSV row is deserialized independently using the column headers as field names. All values start as strings; the deserialization engine applies type coercion for each field.

Pass a context object strategy (e.g. CamelCase()) to apply global field-renaming.

Arguments

::Type{T}: target struct type per row.
x: CSV text as String or Vector{UInt8}.
strategy: optional context object.

Keyword arguments

delimiter::AbstractString = ",": field separator forwarded to parse_csv.

Returns

Vector{T} — one element per data row.

Throws

ParseError: if x is malformed CSV.
MissingFieldError: if a required struct field is absent.
TypeMismatchError: if a column value cannot be coerced to the field type.

Examples

julia> struct Person; name::String; age::Int; end

julia> from_csv(Person, "name,age\nAlice,30\nBob,25")
2-element Vector{Person}:
 Person("Alice", 30)
 Person("Bob", 25)

See also: try_from_csv, to_csv, parse_csv.

source

Serde.SerdeCsv.try_from_csv — Function

try_from_csv(::Type{T}, x; kw...) -> Union{Vector{T}, SerdeError}
try_from_csv(strategy, ::Type{T}, x; kw...) -> Union{Vector{T}, SerdeError}

Like from_csv but returns a SerdeError instead of throwing on failure.

Returns

Vector{T} on success.
A ParseError or DeserError subtype on failure.

Serialization

Serde.SerdeCsv.to_csv — Function

to_csv(data::Vector{T}; delimiter = ",", headers = String[], with_names = true) -> String
to_csv(strategy, data::Vector{T}; delimiter = ",", headers = String[], with_names = true) -> String

Serialize a vector of structs into a CSV string.

The column headers are derived from the struct field names via Serde.ser_name. Nested structs are flattened using _ as a separator (e.g. address_city). Null values (nothing, missing) are written as empty cells.

Pass a context object strategy (e.g. CamelCase()) as the first argument to apply global field-renaming to the column headers.

Arguments

data::Vector{T}: the rows to serialize.
strategy: optional context object.

Keyword arguments

delimiter::String = ",": field separator character.
headers::Vector{String} = String[]: explicit column order. When empty, all columns are included in field-declaration order.
with_names::Bool = true: include a header row as the first line.

Returns

A CSV-formatted String. Returns "" if data is empty.

Examples

julia> struct Person; name::String; age::Int; end

julia> to_csv([Person("Alice", 30), Person("Bob", 25)]) |> print
name,age
Alice,30
Bob,25

julia> to_csv([Person("Alice", 30)]; with_names=false) |> print
Alice,30