data accumulation with pandas

Question

I'd like to accumulate like this.

timestamp	id	strength
1383260400000	1	strength accumulated by square id1 or id2 == 1
1383260400000	2	strength accumulated by square id1 or id2 == 2

if timestamp is different, making another row.

---

here is code. but, result is below.

df2["strength"] = pd.to_numeric(df["strength"])
 for i in range(len(df)):
     if (df.iloc[i - 1, 0] != df.iloc[i, 0]) or \
             df.iloc[i - 1, 1] != df.iloc[i, 1]:
         k = k + 1
         print(k)
     df2.loc[k, "timestamp"] = df.iloc[i, 0]
     df2.loc[k, "id"] = df.iloc[i, 1]
     df2.loc[k, "strength"] = df2.loc[k, "strength"] + df.iloc[i, 3]

timestamp	id	strength
		4.890249263573871e-05
1383260400000	1	0.00692127440351541
1383260400000	2	0.0031095399526760983
		0.0001136447068307
		0.0001557162945607
		0.0002134019939307

---

here is raw input data.

timestamp	squareid1	squareid2	strength
1383260400000	1	751	1.0024511991972666e-06
1383260400000	1	752	4.636474149834957e-05
1383260400000	1	753	0.0001289563565283
1383260400000	1	754	6.849944595888879e-05
1383260400000	1	755	9.220729726979178e-07
1383260400000	2	1	5.1182385382982246e-05
1383260400000	2	55	5.1182385382982246e-05
1383260400000	2	56	3.0469237948520085e-06
1383260400000	2	152	8.513520849740889e-05

Answer 1

df['group_me'] = df.apply(lambda x: x.squareid1 if x.squareid1 - 1 != x.squareid2 else x.squareid2, axis=1)
output = df.groupby(['timestamp', 'group_me'], as_index=False)['strength'].sum()
print(output)

Output:

       timestamp  group_me  strength
0  1383260400000       1.0  0.000297
1  1383260400000       2.0  0.000139