I would like to add to the other comments and describe the process as it applies to general programming logic. In other words whether we are working in an ETL platform generally or in power Bi/ Power apps, we are working in generic ETl terms that can help frame the discussion. This helps for others who may be facing the issue in PowerBi or Power Pivot. As an example I cite the linked in article below (see references).
Data can come from various sources, but in this case let us look at an example with just two sources, csv tables. Power Bi allows us to connect them via foreign key, a process we both know. However, in the attempt to aggregate the data, this might fail if the data types are not correlated. In other words two tables that have same column names, product and price, might not aggregate as one. This is because the "Type" was stored as "text" in one and "integer" in the other. If the items containing "k" or "m" were stored as text, then they cannot play well with integer values. For instance sorting them would result to sorting those without the "k" and placing those with the "k" separately. Let's fix this! As mentioned we would consider cleaning the data first. We wouldn't need to clean it in PowerBi, moreover. We would clean the data in simple ways. In generic programming terms we would think about functions to remove the last bit of a string of items. Trim and Strip come to mind. In the comment above we referenced another method: search and replace. The comment should be clarified to simplify the rationale not compicate it. By searching for and replacing a "k" with a space for example one could replace the k, then trim it to remove extra spacing for format purposes if needed. Power query allows us to add conditional columns. If you are able to understand "M" query language we are simply using expressions to times a column by 1000. To get to that step we would apply the split column by special character, the k. Thus 100k would be two columns, column1.1 100 and column 1.2 holds the k. We would add a conditional column to multiply by 1000 those values in the column 1.1 which had a "K". So we would select column1.1 if it does not end in "m". You would select column1.1 = column1.1 if there are no values other than those which ended in k, for example. In the Formula editor add times 1000 and we did succesfully add the column converting 100k to 100,000.
References: from LinkedIN - https://www.linkedin.com/newsletters/7267597293221031936/