Accuracy 1 on test set

from my Machine Learning course, I know that 100% accuracy is overfitting but I don’t know why that is the case for me.

# Gaussian Naive Bayes
from sklearn.naive_bayes import GaussianNB

gnb = GaussianNB().fit(X_train, y_train)
gnb_predictions = gnb.predict(X_test)
gnb_acc = accuracy_score(y_test, gnb_predictions)
gnb_probs = gnb.predict_proba(X_test)

print(gnb_acc)

I am testing on unseen data. The data was artificially generated and I think it doesn’t have noise. The data has 10k entries but around half are NaNs. Can that be the problem? Or is it due to how I preprocess the data? In the preprocessing, I use LabelEncoder() since I have some strings and some floats in the features. Apart from that, I am splitting like a normal person lol

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.34, random_state=10)

>Solution :

May be three reasons:

1. Small data set (5-10 rows)
2. All targets are equal
3. Very high correlation between features and target

Please, show dataset.