ValueError: Shape of passed values is (1, 1), indices imply (1, 2)

[Haiker-hub]

Hi, I'm a newbie, and while using a directive like "python -m graphrag.index --root./indexing", I get the error described in the title at the "create_base_entity_graph" step. In the code is "cluster_graph.py" in line 101 "output_df[[level_to, to]] = pd.DataFrame(output_df[to].tolist(), index=output_df.index, columns=[level_to, to]). indexing's input folder is the default.txt provided by the source code. How to solve this problem?

[shaoqing404]

这个问题似乎是模型导致的。请注意你的default.txt必须是utf8且不能有特殊的意外字符

[fenzi1223]

我也是这个问题，请问您解决了吗？如何解决呢

[shaoqing404]

我也是这个问题，请问您解决了吗？如何解决呢

更新graphrag到新版本

[pholz]

same issue. based on translating the answers above to english, i looked into 2 things. 1) my files are UTF-8 encoded... and I get the same problem with a single pure ASCII file. 2) updating graphRAG - how? the graphrag dir is severely different from the official graphrag repo.

[shaoqing404]

same issue. based on translating the answers above to english, i looked into 2 things. 1) my files are UTF-8 encoded... and I get the same problem with a single pure ASCII file. 2) updating graphRAG - how? the graphrag dir is severely different from the official graphrag repo.

Track the latest branch