From c25278538c5c594bca99eb53ccd19e105240b20e Mon Sep 17 00:00:00 2001
From: Petr <petr@keboola.com>
Date: Thu, 12 Mar 2026 14:43:23 +0100
Subject: [PATCH] Simplify profiler config: use single SAMPLE_SIZE parameter
 (KISS)

Replace SAMPLE_THRESHOLD + SAMPLE_SIZE with single SAMPLE_SIZE:
- If table > SAMPLE_SIZE: sample that many rows
- Otherwise: use all rows

Cleaner, easier to configure.
---
 src/profiler.py | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/src/profiler.py b/src/profiler.py
index c4cf69a..00e658b 100644
--- a/src/profiler.py
+++ b/src/profiler.py
@@ -34,8 +34,7 @@ logger = logging.getLogger(__name__)
 # ---------------------------------------------------------------------------
 # Profiler configuration
 # ---------------------------------------------------------------------------
-SAMPLE_THRESHOLD = 500_000  # Sample tables larger than this
-SAMPLE_SIZE = 500_000
+SAMPLE_SIZE = 500_000  # If table > this, sample this many rows; else use all
 MAX_CATEGORICAL_DISTINCT = 50  # Treat as categorical if unique <= this
 TOP_VALUES_LIMIT = 10  # Number of top values for categorical columns
 HISTOGRAM_BINS = 15  # Number of bins for numeric histograms
@@ -689,7 +688,7 @@ def profile_table(
 
     # Materialize into temp table — reads parquet files once instead of per-query
     view_name = "tbl"
-    sampled = total_rows > SAMPLE_THRESHOLD
+    sampled = total_rows > SAMPLE_SIZE
     if sampled:
         con.execute(
             f"CREATE TEMP TABLE {view_name} AS SELECT * FROM {read_expr} USING SAMPLE {SAMPLE_SIZE} ROWS"