df 0.2.1

A dataframe for Dart. An abstraction to help manipulating and analysing data

Df

A dataframe for Dart

Usage

Create

From csv:

final df = await DataFrame.fromCsv("dataset/stocks.csv");

Note: the type of the records are infered from the data. The first line of the csv must contains the headers for the column names. Optional parameters:

dateFormat: the string format of the date: reference. Ex: MMM dd yyyy

timestampCol: the column to be treated as a timestamp. Ex: timestamp

timestampFormat: the format of the timestamp: seconds, milliseconds or microseconds. Ex: TimestampFormat.microseconds

verbose: set to true to print some info

From records:

final rows = <Map<String, dynamic>> rows[
   <String, dynamic>{"col1": 21, "col2": "foo", "col3": DateTime.now()},
   <String, dynamic>{"col1": 22, "col2": "bar", "col3": DateTime.now()},
];
final df = DataFrame.fromRows(rows);

Select

final List<Map<String, dynamic>> rows = df.rows;
// select a subset of rows
final List<Map<String, dynamic>> rows = df.subset(0,100);
/// select records for a column
final List<double> values = df.colRecords<double>("col2");
/// select list of records
final List<List<dynamic>> records = df.records;

Mutate

Add data:

// add a row
df.addRow(<String,dynamic>{"col1": 1, "col2": 2.0});
// add a line of records
df.addRecord(<dynamic>[1, 2.0]);

Remove data:

df.removeFirstRow();
df.removeLastRow();
// remove the third row
df.removeRowAt(2);
// limit the dataframe to 100 rows starting from index 30
df.limit(100, startIndex: 30);

Copy a dataframe:

// get a new dataframe from the existing one
final DataFrame df2 = df.copy_();
// get a new dataframe with limited data
final DataFrame df2 = df.limit_(100);

Count

Nulls and zeros:

final int n = df.countNulls_("col1");
final int n = df.countZeros_("col1");

Columns:

final int mean = df.mean("col1");
final int sum = df.sum("col1");
final int max = df.max("col1");
final int min = df.min("col1");

Info

final int numRecords= df.length;
final List<DataFrameColumn> cols = df.columns;
final List<String> colNames = df.columnsNames;
// print info and sample data
df.head();
// like head with a bit more details
df.show();

Conventions

All the dataframe operations are inplace. All the methods that return objects end with an underscore. Example:

// inplace
df.limit(30);
// get a new dataframe with limited data
final DataFrame df2 = df.limit_(30);

Vocabulary conventions:

• A row is a map of key/values pair
• A record is a single cell value
• An index is a row position
• An indice is a column position