```dagster core errors DagsterInvalidDefinitionError In pipe

Question

```dagster core errors DagsterInvalidDefinitionError In pipeline pipe 2 received invalid type lt class str gt for input name at position 1 in solid invocation hello2 Must pass the output from previous solid invocations or inputs to the composition function as inputs when invoking solids during composition ```

Drew Sonne · Answer

I m still learning but from what I understand you can t just use normal strings and values within a ` pipeline` function The functions and variables within a ` pipeline` function describe a dag they re not real functions So any variables need to be passed through the solids One pattern to achieve what you re doing would be to use what I ve seen around as factories eg ``` solid output defs= OutputDefinition int def hello yield Output 1 def hello2 factory name str age int solid input defs= InputDefinition name str InputDefinition age int def hello2 hello int print f Hello name age+hello return hello2 pipeline def pipe 2 hello2 = hello2 factory Marcus 20 x = hello y = hello2 x ```

Darren Ng · Answer

Also having the same error as Alejandro Is this factories method the only way to solve this

Darren Ng · Answer

Very new to dagster here I ve been trying to read a row of params from google sheet using that to query company s database then send some metrics back to write in the same sheet but different cell ``` solid description= Read from google sheet for data pipeline output defs= OutputDefinition name= row data OutputDefinition name= gspread object def read from gsheet gt List print Instantiate gspread sheet object and read a row row = sheet row values 2 yield Output row output name= row data yield Output sheet output name= gspread object solid description= Write to cells in Gsheet def write to gsheet sheet cell str value int sheet update cell value solid description= Read from google sheet and query data from NLP def query NLP data context row received row = row print Sending request to query NLP graph and return metrics some bol metrics = 1000 <http context log info|context log info> f Returned NLP metrics some bol metrics return some bol metrics pipeline def gather pipeline row sheet = read from gsheet some bol metrics = query NLP data row write to gsheet sheet H2 some bol metrics ``` In this example here I m reading a row from a gspread sheet object sending that to another solid which should query company db and return some metrics here 1000 then pass that into another solid which writes to the same sheet object and write into H2 But I m getting ``` home dazza anaconda3 envs versedai lib python3 8 site packages dagster core definitions composition py 65 UserWarning While in pipeline context gather pipeline received an uninvoked solid write to gsheet warnings warn warning message strip home dazza anaconda3 envs versedai lib python3 8 site packages dagster core workspace context py 504 UserWarning Error loading repository location gather pipeline py dagster core errors DagsterInvalidDefinitionError In pipeline gather pipeline received invalid type lt class str gt for input cell at position 1 in solid invocation write to gsheet Must pass the output from previous solid invocations or inputs to the composition function as inputs when invoking solids during composition ```

prha · Answer

Echoing Drew here the pipeline establishes how each solid depends on each other via inputs You could use a solid factory as Drew demonstrates or you could provide the value via config instead of via input

prha · Answer

There are cases where you might prefer an alternate way of resolving inputs see <https docs dagster io concepts io management unconnected inputs> but I suspect that s not the best fit for what you re trying to accomplish here