Skip to content

Commit a8072fb

Browse files
committed
Documentation for Select by Data Index
1 parent e8559e9 commit a8072fb

File tree

5 files changed

+36
-0
lines changed

5 files changed

+36
-0
lines changed

doc/visual-programming/source/index.rst

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -37,6 +37,7 @@ Data
3737
widgets/data/createclass
3838
widgets/data/randomize
3939
widgets/data/concatenate
40+
widgets/data/select-by-data-index
4041
widgets/data/paintdata
4142
widgets/data/pythonscript
4243
widgets/data/featureconstructor
2.34 KB
Loading
204 KB
Loading
8.82 KB
Loading
Lines changed: 35 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,35 @@
1+
Select by Data Index
2+
====================
3+
4+
Match instances by index from data subset.
5+
6+
Inputs
7+
Data
8+
reference data set
9+
Data Subset
10+
subset to match
11+
12+
Outputs
13+
Data
14+
subset from reference data set that matches indices from subset data
15+
16+
17+
**Select by Data Index** enables matching the data by indices. Each row in a data set has an index and given a subset, this widget can match these indices to indices from the reference data. Most often it is used to retrieve the original data from the transformed data (say, from PCA space).
18+
19+
.. figure:: images/Select-by-data-index-stamped.png
20+
:scale: 50%
21+
22+
1. Information on the reference data set. This data is used as index reference.
23+
2. Information on the data subset. The indices of this data set are used to find matching data in the reference data set. Matching data are on the output by default.
24+
25+
Example
26+
-------
27+
28+
A typical use of **Select by Data Index** is to retrieve the original data after a transformation. We will load *iris.tab* data in the **File** widget. Then we will transform this data with **PCA**. We can project the transformed data in a **Scatter Plot**, where we can only see PCA components and not the original features.
29+
30+
Now we will select an interesting subset (we could also select the entire data set). If we observe it in a **Data Table**, we can see that the data is transformed. If we would like to see this data with the original features, we will have to retrieve them with **Select by Data Index**.
31+
32+
Connect the original data and the subset from **Scatter Plot** to **Select by Data Index**. The widget will match the indices of the subset with the indices of the reference (original) data and output the matching reference data. A final inspection in another **Data Table** confirms the data on the output is from the original data space.
33+
34+
.. figure:: images/Select-by-Data-Index-Example1.png
35+

0 commit comments

Comments
 (0)