Your question is a bit confusing. So, as much as I understood from your examples, for each sample of *list_asm*, you want to extract the very first word from the string. The thing you are doing wrong is treating the sting as a list. That is, `['uncomisd xmm2, xmm2', 'jp 0x40', ...]` is considered as a string python python, not a list. Thus, you need to extract the strings from your list first, then you can't take the first words from all these strings. To achieve that, you can use a regular expression to find all the strings that are inside of quotes `'...'`. I used a dummy file *dataF.csv* to imitate your data. import pandas as pd import re # Read the file into dataframe dataFrame = pd.read_json("dataset.json", lines=True) # First extract the strings the take the first word of each string dataFrame['opcodes'] = dataFrame['lista_asm'].apply(lambda x: [i.split()[0] for i in re.findall("'([^']*)'", x)]) print(dataFrame) or modular form of the code would be: import pandas as pd import re # Function to extract the first words from each string def extractFirstWord(str): listOfWords = re.findall("'([^']*)'", str) return [i.split()[0] for i in listOfWords] # Read the file into dataframe dataFrame = pd.read_json("dataset.json", lines=True) dataFrame['opcodes'] = dataFrame['lista_asm'].apply(lambda x: extractFirstWord(x)) print(dataFrame) The result: [![enter image description here][1]][1] [1]: https://i.sstatic.net/sZ7oR.png