Writing CSV files

8.4. Writing CSV files#

To export a TimeSeries to a CSV file, we do the opposite sequence. We first convert the TimeSeries to a DataFrame, then we use the DataFrame’s to_csv method to create a CSV file.

In this example, we read some 3D marker positions from a sample C3D file and export these positions as a CSV file. Let’s first download and read a sample C3D file. This file is provided by https://www.c3d.org as a test suite for C3D software development. For conciseness, we will only keep the following markers: “RFT1” and “RFT2”.

import kineticstoolkit.lab as ktk
import pandas as pd


filename = ktk.doc.download("c3d_test_suite_sample.c3d")
c3d_contents = ktk.read_c3d(filename, convert_point_unit=True)
markers = c3d_contents["Points"].get_subset(["RFT1", "RFT2"])

markers

TimeSeries with attributes:
         time: <array of shape (450,)>
         data: <dict with 2 entries>
    time_info: {'Unit': 's'}
    data_info: {'RFT1': {'Unit': 'm'}, 'RFT2': {'Unit': 'm'}}
       events: []

markers.data

{
    'RFT1': <array of shape (450, 4)>
    'RFT2': <array of shape (450, 4)>
}

To convert this TimeSeries to a CSV file, we first create a DataFrame:

df = markers.to_dataframe()

df
RFT1[0] RFT1[1] RFT1[2] RFT1[3] RFT2[0] RFT2[1] RFT2[2] RFT2[3]
0.00 0.248583 0.226833 0.037417 1.0 0.212667 0.218333 0.088917 1.0
0.02 0.249000 0.226750 0.037000 1.0 0.213167 0.218500 0.088250 1.0
0.04 0.249333 0.226333 0.036667 1.0 0.213333 0.216917 0.088250 1.0
0.06 0.248833 0.226667 0.036583 1.0 0.213000 0.216667 0.088167 1.0
0.08 0.249917 0.226250 0.036833 1.0 0.214167 0.216417 0.087833 1.0
... ... ... ... ... ... ... ... ...
8.90 0.325083 2.243583 0.034333 1.0 0.294083 2.237917 0.086333 1.0
8.92 0.323583 2.249417 0.032583 1.0 0.293250 2.238667 0.086167 1.0
8.94 0.324250 2.247750 0.033250 1.0 0.290917 2.238750 0.086833 1.0
8.96 0.325250 2.247833 0.034000 1.0 0.291750 2.239000 0.086250 1.0
8.98 0.324583 2.248000 0.033750 1.0 0.291000 2.238750 0.087083 1.0

450 rows × 8 columns

Then we export this DataFrame to a CSV file. Let’s print the first 3 lines of this file:

df.to_csv("output.csv", index_label="Time")

Here are the first lines of the exported CSV file, ready to be processed in another software:

Time,RFT1[0],RFT1[1],RFT1[2],RFT1[3],RFT2[0],RFT2[1],RFT2[2],RFT2[3]
0.0,0.24858334074169397,0.22683334009349346,0.03741666778177023,1.0,0.21266667300462724,0.21833333984017372,0.08891666931658983,1.0
0.02,0.24900000742077827,0.22675000675767662,0.03700000110268593,1.0,0.21316667301952838,0.21850000651180745,0.08825000263005495,1.0
0.04,0.24933334076404573,0.2263333400785923,0.036666667759418486,1.0,0.2133333396911621,0.2169166731312871,0.08825000263005495,1.0
0.06,0.24883334074914457,0.22666667342185975,0.03658333442360163,1.0,0.21300000634789468,0.21666667312383653,0.0881666692942381,1.0