google cloud plateform Auto ML

Question

Hi guys i'm trying to import a table in "AutoML Tables" , just for a test I created a table with random numbers so don't mind it. Everytime I test to try to import a table it seems not to work It doesn't seems anything to be wrong in my CSV. Only alphanumeric character, no empty cell, just numeric data for the test 3 columns, 300 rows ... But no it isn't working and its telling me :

Error Messages: Invalid column names:

My columns name are : rooms, or , price . I don't know what could be wrong with those names...

thank you for your help

enter image description here

I tried with and without the quote mark

"id","rooms","OR","price","space","toilets"
0,5,8,200,200,1
1,5,8,200,200,1
2,5,8,200,200,1
3,5,8,200,200,1
4,5,8,200,200,1
5,5,8,200,200,1
6,5,8,200,200,1
7,5,8,200,200,1
8,5,8,200,200,1
9,5,8,200,200,1
10,5,8,200,200,1
11,5,8,200,200,1
12,7,9,300,400,2
13,7,9,300,400,2
14,7,9,300,400,2
15,7,9,300,400,2

Could it be that it detects 4 columns and the first one with empty name? — Temu, May 28 '20 at 07:22
no its something else, because even when i feel the blank it does'tn work... — Helena Rousseaux, Jun 01 '20 at 10:32
Helena, would you be so kind as to pass me a example txt in order to reproduce the issue? — Temu, Jun 01 '20 at 14:19
Hi thank you Temu, I don't know how to attached a file here but I paste the data in the question, actually the value doesn't really matter as Its just for a test — Helena Rousseaux, Jun 02 '20 at 08:17
I think I've spotted what might be occurring. I also got an error with the example and rereading your description you mention that you only use 3 columns with 300 rows " > Error Messages: > Too few rows: 16. Minimum number is: 1000". Could you try with at least 1000 rows and let me know what happens? — Temu, Jul 14 '20 at 16:12
I have got the same error, and no of rows is not the issue. I have total of 21K rows. — Muhammad Hassan, Oct 20 '20 at 08:36

score 2 · Accepted Answer · answered Oct 30 '20 at 22:26

I recreated the issue when the CSV was created with an index column.

import pandas as pd

columns = ["id","rooms","OR","price","space","toilets"]
data = [
    [0,5,8,200,200,1],
    [1,5,8,200,200,1],
    [2,5,8,200,200,1],
    [3,5,8,200,200,1],
    [4,5,8,200,200,1],
    [5,5,8,200,200,1],
    [6,5,8,200,200,1],
    [7,5,8,200,200,1],
    [8,5,8,200,200,1],
    [9,5,8,200,200,1],
    [10,5,8,200,200,1],
    [11,5,8,200,200,1],
    [12,7,9,300,400,2],
    [13,7,9,300,400,2],
    [14,7,9,300,400,2],
    [15,7,9,300,400,2]
]

df = pd.DataFrame(data=data, columns=columns)
# resampled the data to avoid AutoMLTables error: 
# Too few rows: 16. Minimum number is: 1000
df = df.sample(1000, replace=True)
df.to_csv('/your/data/path/here', index=True)

But if I set index to False and recreated the file the import succeeds.

It's also possible that an earlier version of AutoMLTables didn't like that you used a reserved keyword, id, as a column name but this is unlikely.

google cloud plateform Auto ML

1 Answers1