How to generate data or sample data from a summary tables? I have a dataset like this:

I only have summary tables of the "loneliness score" for overall, for each sex, each age group, and sex * age interactions. Besides this, I have 23 items that are also collected from the same sample with tables like the one I have provided.
My goal is to generate an dataset of N = 2,000 which follows the summary data's structure...
In this case, which statistical methods could be applied?
Thanks in advance for your help!
 A: The table gives you all the data, just with binned age. The table says that there were 13 males aged 16-19 who answered "always" to the question, so you can re-create the original data just by repeating the row
male, 16-19, always

thirteen times in your CSV file. If you'd like to downsample it, just pick the random rows from the CSV generated as above or do a lottery where each cell gets sampled randomly, with the probability proportional to the counts, i.e. 13 / 11867 in the case of the cell mentioned above.
However, you say that you have many such tables, so I'm worried that you would like to "generate" the data from each table and combine them, and that won't work. If you generate the tables independently of each other, you are ignoring all the correlations between the variables, so calculating any multivariate statistics won't make sense anymore. Before calculating the statistics you would know that the variables in your generated data are uncorrelated, the regression parameters are close to zero, etc because of how the data was created.
To give you an example, let's use the same cell as above. Say that you don't have the interaction of gender and age rows in your data, just the age, and gender alone. With the interaction, the fraction of the males aged 16-19 who answered "always" is 13 / 11867 = 0.001, while for both genders together this is 379 / 11867 = 0.032 and for all the males 5476 / 11867 = 0.46, so if you considered those variables as independent the 0.032 * 0.46 = 0.015, that is a way off comparing to the count in the actual data that was 0.001.
So from this table, you can re-create this particular data, but there is no way to combine it with other tables.
