Reading and Writing CSV files

Arrow provides a fast CSV reader allowing ingestion of external data as Arrow tables.

Basic usage

A CSV file is read from a InputStream.

#include "arrow/csv/api.h"

{
   // ...
   arrow::io::IOContext io_context = arrow::io::default_io_context();
   std::shared_ptr<arrow::io::InputStream> input = ...;

   auto read_options = arrow::csv::ReadOptions::Defaults();
   auto parse_options = arrow::csv::ParseOptions::Defaults();
   auto convert_options = arrow::csv::ConvertOptions::Defaults();

   // Instantiate TableReader from input stream and options
   auto maybe_reader =
     arrow::csv::TableReader::Make(io_context,
                                   input,
                                   read_options,
                                   parse_options,
                                   convert_options);
   if (!maybe_reader.ok()) {
      // Handle TableReader instantiation error...
   }
   std::shared_ptr<arrow::csv::TableReader> reader = *maybe_reader;

   // Read table from CSV file
   auto maybe_table = reader->Read();
   if (!maybe_table.ok()) {
      // Handle CSV read error
      // (for example a CSV syntax error or failed type conversion)
   }
   std::shared_ptr<arrow::Table> table = *maybe_table;
}

A CSV file is written to a OutputStream.

#include <arrow/csv/api.h>
{
    // Oneshot write
    // ...
    std::shared_ptr<arrow::io::OutputStream> output = ...;
    auto write_options = arrow::csv::WriteOptions::Defaults();
    if (WriteCSV(table, write_options, output.get()).ok()) {
        // Handle writer error...
    }
}
{
    // Write incrementally
    // ...
    std::shared_ptr<arrow::io::OutputStream> output = ...;
    auto write_options = arrow::csv::WriteOptions::Defaults();
    auto maybe_writer = arrow::csv::MakeCSVWriter(output, schema, write_options);
    if (!maybe_writer.ok()) {
        // Handle writer instantiation error...
    }
    std::shared_ptr<arrow::ipc::RecordBatchWriter> writer = *maybe_writer;

    // Write batches...
    if (!writer->WriteRecordBatch(*batch).ok()) {
        // Handle write error...
    }

    if (!writer->Close().ok()) {
        // Handle close error...
    }
    if (!output->Close().ok()) {
        // Handle file close error...
    }
}

Note

The writer does not yet support all Arrow types.

Column names

There are three possible ways to infer column names from the CSV file:

  • By default, the column names are read from the first row in the CSV file

  • If ReadOptions::column_names is set, it forces the column names in the table to these values (the first row in the CSV file is read as data)

  • If ReadOptions::autogenerate_column_names is true, column names will be autogenerated with the pattern “f0”, “f1”… (the first row in the CSV file is read as data)

Column selection

By default, Arrow reads all columns in the CSV file. You can narrow the selection of columns with the ConvertOptions::include_columns option. If some columns in ConvertOptions::include_columns are missing from the CSV file, an error will be emitted unless ConvertOptions::include_missing_columns is true, in which case the missing columns are assumed to contain all-null values.

Interaction with column names

If both ReadOptions::column_names and ConvertOptions::include_columns are specified, the ReadOptions::column_names are assumed to map to CSV columns, and ConvertOptions::include_columns is a subset of those column names that will part of the Arrow Table.

Data types

By default, the CSV reader infers the most appropriate data type for each column. Type inference considers the following data types, in order:

It is possible to override type inference for select columns by setting the ConvertOptions::column_types option. Explicit data types can be chosen from the following list:

  • Null

  • All Integer types

  • Float32 and Float64

  • Decimal128

  • Boolean

  • Date32 and Date64

  • Time32 and Time64

  • Timestamp

  • Binary and Large Binary

  • String and Large String (with optional UTF8 input validation)

  • Fixed-Size Binary

  • Dictionary with index type Int32 and value type one of the following: Binary, String, LargeBinary, LargeString, Int32, UInt32, Int64, UInt64, Float32, Float64, Decimal128

Other data types do not support conversion from CSV values and will error out.

Dictionary inference

If type inference is enabled and ConvertOptions::auto_dict_encode is true, the CSV reader first tries to convert string-like columns to a dictionary-encoded string-like array. It switches to a plain string-like array when the threshold in ConvertOptions::auto_dict_max_cardinality is reached.

Nulls

Null values are recognized from the spellings stored in ConvertOptions::null_values. The ConvertOptions::Defaults() factory method will initialize a number of conventional null spellings such as N/A.

Character encoding

CSV files are expected to be encoded in UTF8. However, non-UTF8 data is accepted for Binary columns.

Write Options

The format of written CSV files can be customized via WriteOptions. Currently few options are available; more will be added in future releases.

Performance

By default, the CSV reader will parallelize reads in order to exploit all CPU cores on your machine. You can change this setting in ReadOptions::use_threads. A reasonable expectation is at least 100 MB/s per core on a performant desktop or laptop computer (measured in source CSV bytes, not target Arrow data bytes).