ValueError: could not convert string to float — how to fix

ValueError: could not convert string to float: ''
ValueError: could not convert string to float: '1,234.50'

Quick fix: Look at the quoted value in the message. Clean it before converting: float(s.strip().replace(',', '')), skip empty strings, or in pandas use pd.to_numeric(df['col'], errors='coerce') to turn bad values into NaN.

What it means

float(x) accepts strings like "3.14", " 42 ", "1e-5", "inf" and "nan". Anything else raises this error, and the message shows the repr of the bad string:

float("")          # could not convert string to float: ''
float("1,234.5")   # could not convert string to float: '1,234.5'
float("$9.99")     # could not convert string to float: '$9.99'
float("3,5")       # European decimal comma

Read the quoted part carefully: '12.5\xa0' (non-breaking space) or '12.5\n' hidden inside data look like valid numbers when printed normally.

Cause 1: empty strings

Blank CSV cells, empty form fields, or a trailing newline split into an empty line.

value = row["price"].strip()
price = float(value) if value else None

Cause 2: thousands separators, currency, units, percent

import re

def parse_number(s: str) -> float:
    s = s.strip().replace(",", "")      # 1,234.50 -> 1234.50
    s = re.sub(r"[^\d.eE+-]", "", s)    # drop $, %, kg, spaces
    return float(s)

parse_number("$1,234.50")   # 1234.5
parse_number("45 %")        # 45.0

Be careful: stripping characters blindly can hide genuinely bad data. Log values you had to clean.

Cause 3: decimal comma (locale formats)

"3,5" means 3.5 in much of Europe. Either replace the comma:

float("3,5".replace(",", "."))

or, for mixed formats like "1.234,56", remove the dots first, then replace the comma. When reading CSVs with pandas, tell it the format:

pd.read_csv("data.csv", decimal=",", thousands=".")

Cause 4: header row or text in a numeric column

Reading a CSV manually and converting the header line ("price") fails. Skip it:

import csv
with open("data.csv", newline="") as f:
    for row in csv.DictReader(f):
        total += float(row["price"])

Fix in pandas

df["col"].astype(float) fails on the first bad value. To find and handle them:

import pandas as pd

nums = pd.to_numeric(df["price"], errors="coerce")   # bad values -> NaN
print(df.loc[nums.isna() & df["price"].notna(), "price"].unique())  # what failed
df["price"] = nums

To clean common formatting first:

df["price"] = pd.to_numeric(
    df["price"].astype(str).str.replace(r"[$,\s]", "", regex=True),
    errors="coerce",
)

Fix for user input

Wrap the conversion and ask again:

while True:
    try:
        amount = float(input("Amount: "))
        break
    except ValueError:
        print("Please enter a number like 12.5")

Related: int()

int("3.7") fails with invalid literal for int() with base 10. Use int(float("3.7")) if you want to truncate, or round(float(s)).

Last updated 2026-11-04.