Filter

Filter Table

Filters rows in a TableData object using a pandas query string.

Details

Examples:

• Area > 100 — keep rows where Area is greater than 100
• Area > 50 and Circularity > 0.8 — multiple conditions
• Group == "Control" — match a specific text value
• Group != "Background" — exclude rows
• Area > Area.mean() — compare to column statistics
• label in [1, 2, 5] — match specific values from a list

Uses pandas DataFrame.query() syntax. Column names with spaces need backticks: `Column Name` > 10

Direction	Port	Type
Input	in	table
Output	out	table

---

Top N

Extracts the top (or bottom) N rows ranked by a numeric column.

Details

Outputs:

• top_n — the selected N rows
• rest — all remaining rows not in top_n

Parameters:

• Rank By Column — numeric column to rank by
• N — number of rows to select
• Select — Top (largest) or Bottom (smallest)

Direction	Port	Type
Input	in	table
Output	top_n	table
Output	rest	table

Properties: N (number of rows), Select

---

Column Value Split

Splits a table into two outputs based on whether a column's value matches a list of specified values.

Details

• Values — comma-separated. * anywhere triggers glob matching:
• Control* — starts with "Control"
• *treated — ends with "treated"
• *GFP* — contains "GFP"
• Entries without * are exact matches

Outputs:

• matched — rows where the column value matches any entry
• rest — all other rows

Direction	Port	Type
Input	in	table
Output	matched	table
Output	rest	table

Properties: ``

---

Random Sample

Randomly samples N rows from the input table.

Details

If N exceeds the table size, the full table is returned (no error).

Parameters:

• N — number of rows to draw
• Seed — random seed for reproducibility; leave at -1 for a different sample each run

Outputs:

• sampled — the N randomly selected rows
• rest — all remaining rows not in the sample

Direction	Port	Type
Input	in	table
Output	sampled	table
Output	rest	table

Properties: N (rows to sample), Random Seed (-1=random)

---

Drop Duplicates

Removes duplicate rows from a table.

Details

Parameters:

• Subset Columns — comma-separated columns to consider when checking for duplicates. Leave empty to compare all columns.
• Keep — which duplicate to keep: first occurrence, last, or none

Outputs:

• unique — rows after removing duplicates
• dropped — the removed duplicate rows

Direction	Port	Type
Input	in	table
Output	unique	table
Output	dropped	table

Properties: Keep

---