From 5d4c2bf4fc75715fb079f3b5c7d33e91cdad3202 Mon Sep 17 00:00:00 2001
From: ishan <ishan.gupta@talentmonk.com>
Date: Sat, 16 Nov 2024 18:49:13 +0530
Subject: [PATCH] added countDistinct aggregation

---
 src/danfojs-base/aggregators/groupby.ts | 13 ++++++++++++-
 src/danfojs-base/core/series.ts         |  5 +++++
 src/danfojs-base/shared/types.ts        |  2 ++
 3 files changed, 19 insertions(+), 1 deletion(-)

diff --git a/src/danfojs-base/aggregators/groupby.ts b/src/danfojs-base/aggregators/groupby.ts
index 63671914..c933698f 100644
--- a/src/danfojs-base/aggregators/groupby.ts
+++ b/src/danfojs-base/aggregators/groupby.ts
@@ -231,7 +231,7 @@ export default class Groupby {
   private arithemetic(operation: {[key: string] : Array<string> | string} | string): { [key: string ]: {} } {
 
     const opsName = [ "mean", "sum", "count", "mode", "std", "var", "cumsum", "cumprod",
-    "cummax", "cummin", "median" , "min", "max"];
+    "cummax", "cummin", "median" , "min", "max", "countdistinct"];
     if (typeof operation === "string" ) {
       if (!opsName.includes(operation)) {
         throw new Error(`group operation: ${operation} is not valid`)
@@ -377,6 +377,9 @@ export default class Groupby {
           return sum
         }, 1)
         break;
+      case "countdistinct":
+        data.push(new Set(colVal).size);
+        break;
     }
     return data
   }
@@ -516,6 +519,14 @@ export default class Groupby {
     return this.operations("min")
   }
 
+  /**
+   * Obtain the distinct number of columns for each group
+   * @returns DataFrame
+   */
+  countDistinct(): DataFrame{
+    return this.operations("countdistinct")
+  }
+
   /**
    * Obtain a specific group
    * @param keys Array<string | number>
diff --git a/src/danfojs-base/core/series.ts b/src/danfojs-base/core/series.ts
index e42f169d..60092d49 100644
--- a/src/danfojs-base/core/series.ts
+++ b/src/danfojs-base/core/series.ts
@@ -1210,6 +1210,11 @@ export default class Series extends NDframe implements SeriesInterface {
         return this.cumOps("min", ops);
     }
 
+    countDistinct(options?: { inplace?: boolean; }): Series | void {
+        const ops = { inplace: false, ...options }
+        return this.cumOps("countdistinct", ops);
+    }
+
 
     /**
        * Returns cumulative maximum over a Series
diff --git a/src/danfojs-base/shared/types.ts b/src/danfojs-base/shared/types.ts
index d89be170..617b2fcc 100644
--- a/src/danfojs-base/shared/types.ts
+++ b/src/danfojs-base/shared/types.ts
@@ -154,6 +154,7 @@ export interface SeriesInterface extends NDframeInterface {
     cumMin(options?: { inplace?: boolean }): Series | void
     cumMax(options?: { inplace?: boolean }): Series | void
     cumProd(options?: { inplace?: boolean }): Series | void
+    countDistinct(options?: { inplace?: boolean }): Series | void
     lt(other: Series | number | Array<number> | boolean[]): Series
     gt(other: Series | number | Array<number> | boolean[]): Series
     le(other: Series | number | Array<number> | boolean[]): Series
@@ -233,6 +234,7 @@ export interface DataFrameInterface extends NDframeInterface {
     cumMin(options?: { axis?: 0 | 1 }): DataFrame | void
     cumMax(options?: { axis?: 0 | 1 }): DataFrame | void
     cumProd(options?: { axis?: 0 | 1 }): DataFrame | void
+    countDistinct(options?: { axis?: 0 | 1 }): DataFrame | void
     copy(): DataFrame
     resetIndex(options: { inplace?: boolean }): DataFrame | void
     setIndex(