在制作LaTex表格的时候,常常需要合并行和列,本文介绍的是使用 multirow 包里面的命令完成合并行列的方法。
需要先导入以下库:
usepackage{multirow}
举例:
egin{table*}
%usepackage{multirow}
center
egin{tabular}{|c|c|c|c|c|}
hline % horizontal line
multirow{2}{*}{1 & 2} & multicolumn{2}{l|}{ 3 & 4 } & 5 & 6\ %end line
cline{2-5} %short partial horizontal lines from column 2 to column 5
& 7 & 8 & 9 & 10\ %first cell is occupied by the multirow
hline
hline
$ s_{1} $ & 2 & 3 & 4 & 5\
$ s_{2} $ & 2 & 3 & 4 & 5\
hline
hline
$ s_{3} $ & 2 & 3 & 4 & 5\
$ s_{4} $ & 2 & 3 & 4 & 5\
hline
end{tabular}
end{table*}
效果

egin{table}[h]
caption{Optimization summary.}
label{tab:summary}
egin{tabular}{cccc}
oprule
multirow{2}{*}{Design} & multicolumn{2}{c}{Problem size} & multirow{2}{*}{CPU time(s)} \
cline{2-3}
& Binary variables & Constraints\
midrule
OM1 & 199 & 13,017 & 000 \
OM2 & 199 & 1,525 & 000 \
ottomrule
end{tabular}
label{tab:my_label}
end{table}
效果(后面有改进版)

举例:
egin{table}[h]
caption{Optimization summary.}
centering
vspace{2mm}
egin{tabular}{ccccccccc}
oprule
multirow{2}{*}{Model Name} & multicolumn{4}{c}{Building Roof} & multicolumn{4}{c}{Building}\
cline{2-9}
& Precision & Recall & $F_1$ & IoU & Precision & Recall & $F_1$ & IoU\
midrule
FCN-8s cite{long2015fully} & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83\
U-Net cite{ronneberger2015u} & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83\
Res-U-Net cite{xu2018building} & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83\
SegNet cite{badrinarayanan2017segnet} & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83\
DeepLabV3+ cite{chen2018encoder} & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83\
DeConvNet cite{noh2015learning} & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83\
Our Model & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83\
ottomrule
end{tabular}
label{tab:my_label}
end{table}
效果:

上面表格的问题在于两行标题栏距离太近,通过下面解决,也可以参考网站 https://www.tablesgenerator.com (Booktabs table style)
修改代码:
% Please add the following required packages to your document preamble:
% usepackage{booktabs}
% usepackage{multirow}
egin{table}[h]
caption{Optimization summary.}
label{tab:summary}
egin{tabular}{cccc}
oprule
multirow{2}{*}{Design} & multicolumn{2}{c}{Problem size} & multirow{2}{*}{CPU time(s)} \
cmidrule{2-3}
& Binary variables & Constraints\
midrule
OM1 & 199 & 13,017 & 78.7 \
OM2 & 199 & 1,525 & 13.8 \
ottomrule
end{tabular}
label{tab:my_label}
end{table}
效果:(更协调了一些,目前最优)
- toprule:最上面的线(粗)
- cmidrule:标题中间的线(细)
- midrule:中间的线(中)
- bottomrule:最下面的线(粗)

添加表格注释,参考:Latex如何在表格下方显示注脚_小白皮皮-程序员宅基地_latex 表格下方注释
导入以下库:
usepackage{threeparttable}
举例:
egin{table}[h]
centering
caption{Quantitative comparison of four conventionally used metrics obtained from the segmentation results by FCN-8s, U-Net, Res-U-Net, SegNet, DeepLabV3+, DeConvNet, and the proposed model, where the values in bold format are the highest numbers for corresponding metrics.}
vspace{2mm}
egin{threeparttable} % 增加部分
egin{tabular}{ccccccccc}
oprule
multirow{2}{*}{Model Name} & multicolumn{4}{c}{Building Roof} & multicolumn{4}{c}{Building}\
cline{2-9}
& Precision & Recall & $F_1$ & IoU & Precision & Recall & $F_1$ & IoU\
midrule
FCN-8s cite{long2015fully} & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83\
U-Net cite{ronneberger2015u} & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83\
Res-U-Net cite{xu2018building} & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83\
SegNet cite{badrinarayanan2017segnet} & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83\
DeepLabV3+ cite{chen2018encoder} & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83\
DeConvNet cite{noh2015learning} & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83\
Our Model & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83 & 69.83\
ottomrule
end{tabular}
label{tab:my_label}
egin{tablenotes} %增加部分
item The bolded numbers indicate the largest number in the column, easier to find out which one performs better in this format.
end{tablenotes} %增加部分
end{threeparttable} %增加部分
end{table}
效果:
