You can mitigate this error without having to change your data or remove duplicates. Just create a new index with DataFrame.reset_index:

df = df.reset_index()

The old index is kept as a column in your dataframe, but if you don’t need it you can do:

df = df.reset_index(drop=True)

Some prefer:

df.reset_index(inplace=True, drop=True)

best solution from this page:
https://pandas.pydata.org/pandas-docs/version/0.20/merging.html

df = pd.concat([df1, df2], axis=1, join_axes=[df1.index])

This post comes up top when you search for the error but the answers are not complete, so let me add mine. There is another reason this error can happen: If you have duplicate columns in your data frames, you will not be able to concatenate and raise this. In fact, even in the original question there are two columns called timestamp. So it will be better to check if len(df.columns) == len(set(df.columns)) for all the data frames you are trying to concatenate.

As a complement of Nicholas Morley’s answer, when you find even this not works:

df = df.reset_index(drop=True)

You should check whether the columns are unique. When they are not, even reseting index not works. Duplicated columns should be removed first to make it works.

Duplicated column names!

In my case the problem was because I had duplicated column names.

This happens also when you have duplicates in the columns names.

Same Indices Between the Two DFs

Another reason for this issue might be that df1 and df2 might have the same indices, between each other. For example, both the dfs might have the same index idx1.

To check if this is the issue, you can see if the following outputs not an empty list:

print([org_name for org_name in cum_df.index if org_name in df_from_2002.index])

My suggested solution then would be to rename the indices (so df1 would keep having idx1 and you would change idx1 to idx2 in df2) and after concatenating (df1 = pd.concat([df1, df2])), combine the two indices (in case you need to get the sum of them) with this code:

df1.iloc[idx1] = df1.iloc[[idx1, idx2]].sum()

and then remove idx2:

df1.drop([idx2], inplace=True)

This happened to me when I was trying to concat two dataframes that have duplicated column names!

Let’s say that I want to remove the first duplicated column:

duplicated_column = 'column'

df_tmp = df[duplicated_column].T
df_tmp = df_tmp.iloc[1: , :]

df = df.drop([duplicated_column], axis=1)
df = pd.concat([df, df_tmp.T], axis=1)

Answers here helped but concat worked fine for me in some cases even where duplicate columns were present. However, in some cases it didn’t work and raised the InvalidIndexError.

It turned out that it works fine if order of duplicate columns is same but raises an error if order of duplicate columns is different.

Example where it works fine:

df = pd.DataFrame({'a': [1, 2, 3], 'b': [5, 6, 7], 'c': [9, 10, 11]})
df1 = pd.DataFrame({'a': [12], 'b': [13], 'c': [14]})
df.rename(columns={
    'c': 'b'
}, inplace=True)
df1.rename(columns={
    'c': 'b'
}, inplace=True)
print(pd.concat([df, df1]))

Output:
    a   b   b
0   1   5   9
1   2   6  10
2   3   7  11
0  12  13  14

Example where it doesn’t work:

df = pd.DataFrame({'b': [1, 2, 3], 'a': [5, 6, 7], 'c': [9, 10, 11]})
df1 = pd.DataFrame({'a': [12], 'b': [13], 'c': [14]})
df.rename(columns={
    'c': 'b'
}, inplace=True)
df1.rename(columns={
    'c': 'b'
}, inplace=True)
print(pd.concat([df, df1]))

Output:
pandas.errors.InvalidIndexError: Reindexing only valid with uniquely 
valued Index objects

This is because you have duplicated columns. Before concatenating drop duplicated columns in each DataFrame as follows:

df = df.loc[:,~df.columns.duplicated()].reset_index(drop=True)

The problem for me was duplicate column labels, just as many others here mentioned it.
To keep only the first column for duplicates I used below:

df=df.T[~df.T.index.duplicated(keep='first')].T

NOTE: I cannot comment due to lack of status here so adding my comments as answers

What if you want those duplicate columns and still need to proceed ahead? What would someone use?

I have two DFs.
df1 —> # Date Time Ticker Open High Low Close Volume OpenInterest
df2 —> # Date Time Ticker Open High Low Close Volume OpenInterest

df1 holds futures data
df2 holds options data

Both are for same underlying.
df_merged is my dataframe where I want to merge df1 and df2

First, I will add df1 to df_merged

df_merged=pd.concat([df_merged,df1])

Then I will loop through my main df and extract option contracts data, one by one and will merge with df_merged.

for con in av_options:
    print("Contract CON:",con)
    df_op=r_data[r_data["Ticker"]==con]
    name=con.split(".")[0]
    df_op.drop(columns=["Ticker","Date"],inplace=True)
            
    df_op.columns= ["Time",name+"_O",name+"_H",name+ .... likewise
    df_merged=df_merged.merge(df_op,how="outer",on="Time")


                                                                       

This whole process will be repeated as per the users’ entered end date. So if start date to End date have 4 days then we will process this 4 times.

NOTE: We access daily data stored in CSV format. All that code is skipped here.

When I run the code, sometimes I do not get any error. Sometime I have this error.